‘ ".rv‘om‘uwh -

 

 

 

 

 

 

 

71‘: .

mix}.

JK‘fi 52‘3"!

m-J.‘

\q

 

‘.

8:9

1.4:-

 

3'1"»?

v: Ln.

 

 

 

1'

mu‘i

   

 

‘1. an»?

-1

5-11-2-

i}.

“A(' H:

1

395‘!

1 .
“'3; 1':

1.. ..1:~"1};’3§

 

 

u-fln',»
' 14..

‘. .‘uu‘v

       

.. 1.123%...
.‘9‘ 1.. "P 121:1. 1

 

 

 

~ .-.- ... ' .\.
.111 1’11‘1-wui . .
“"923‘1‘ A; ,.
t“‘31”,§1fij‘:ig‘ 1 59:3.
”1.
1'::1‘fi

 

‘1 :ul v ’1’
”WWW... I:
. - ‘1 1. ~
3.:11tx'13u
l" 1 —-~§ 1.1
43‘

u n \
' ' 1 _ ‘.’”;.7"'.‘ 1 ~15.
. . . ~ v 11 7 ~1 . m 1.
. .r..1...1 .. 1 . .1 1 1 , .11- 1 3: 1
4 —1 9.91m . ,- . w r .‘k 11-.5 A 1‘.
-. )3... . , . . . -L» 1 .T ‘71”
. f 1; a: ‘ A . ml .. .{ .'..1g:'32,1“'"._.2.7.3$.,.
' ‘ . "7’33“.“ . 1,. 1. 11.13;“.
‘13"..- ‘ .m.
1

   

 

'33:?

11;-

 

.
1-111 a
.....~.;:~..c:: 1
I“ “#1:“ 'AxMV‘L
~ 123‘

 

.-.-:~>_. . A . »
’1“ 4611 .u‘
kusr . 2:2.

“10117.“ .

;.‘.‘f'.‘\,.,.. ..

i:

1 .
3‘31}; w’w-m‘h
1.1.1.1

w . w “Lafifig‘i :"

M1311?" 421:1 ‘ 1'?»

 

 

 

. - .~ =1.» :w
«‘r. I“ Jig: rfigi‘...‘ thrill
“ iii-f -.., 35.3.2.3: fi::~:;
‘ 1,1 13‘". :11: 13""
(NYV ‘ I: “hi: ‘ Mm" M ‘1"?
6“ ’ '11 ..

' wkvt 1 '7
“111%. 51...:
.1111“...

_ V
,-.1¢.Pl' - iii; ‘1’ L {$33.}. ;;;fi" Wham-41' *.
” .J: 1 n,- .H . s y _ ,1“ 01-1111: in 1.1312}??? s;- «gt-"111.9. 11:1: 73L,
,. v... ( A .3 -= ~ ..:x-5z= “ a .5 1,“ ‘x. E..- ’33; :55... .~*-::£: . mm. “3:“
-",>..:"153‘”3~v'-' .z.-_.. ., ‘ ...:*.£w‘fi« P x 7' E ‘aqgfik "" mfirwfi‘w
"1:11.392‘13‘m v-I‘ntru-f-‘A ‘5“?2'1.‘ J *3“ “(ii-r 1.1.1:: . . . ~ '_ 1 1-1! “5.11.13“. 11 um .
"Q'mfl'f .11. “." 3:1 1 1 1 u 1 {130'1‘1- 1 ‘ “1:11:01
1

 

 
 

'-1 .1
r
1

 

“ 1n {61.3337
my 1. . -. . éfi“ ‘ 1119‘- ;‘A'. .. . ' 1 Izwafgf’ 1‘;
"1.311;. ,5 .‘ztx‘ffiwar. » .w‘-1-cw my“? ' -' ‘ “:1 32:21.1:
«Sift "~21: _ ,r-z 3'31}ng 1 ‘ 551.... .. '-
Eta-1.12:2,“ 11.13.. ‘ "
5 .rm tux-L“, .‘Cfifgm

.1

 

:;n .11-

‘u "' -frc‘. . -33:
SM" 1,;- 1 all]; x - 1'- ' {vamr-«I‘r‘u l
.1-

4‘
1"..2'. «1’ x .11 1"!»-
, 1 . 7‘ .1 . ‘ «gym;
‘13: a v1." ' I: ‘f ‘ fi. ‘ f'

 

'11 “aw I
”at. . 3:1 rung-5V1
s‘" M. 1‘." «3-

 

1 I
.111- 'm.

. ." v.1 5:
9..

‘ :i 1 .
1. . . {fur-1r
, .... .- . .3-
1..’!:,:,',,, ) 1,, . . vrqtgv
., .1-

 

1:.1

.:
.v 0"":
{ 4

, .
, .,
,~'..:'_'-.....-4. _ . , . ...;.:, 1“,.
"."“’d.'" _ . .1 .',,. ,5. H,

 

I

 

 

   

 

 

PM I

Hill “

 

 

J lliiilli'lllillilllllllllflfl'

3 1293 00784 5997

 

ll LIBRARY
Mlchlgan State
, Unlverslty

 

 

 

This is to certify that the

thesis entitled

Analysis Of Neural Network Response
with Varied Neuron Models and Interconnection Patterns

presented by

David Barnard Pierce

has been accepted towards fulfillment
of the requirements for

 

 

 

' 0
Master 3 degree in Electrical
Engineering
I/A (J? l: I“ L 1.1:" 7/1» [Hf/.2111 Yul—‘—
' Major professor
13 Nov 1991

Date

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

 

 

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fifi

 

 

 

A
MSU Is An Affirmative ActiorVEqueI Opportunity Institution
czlcircmpma-pn

 

ANALYSIS OF
NEURAL NETWORK RESPONSE
WITH VARIED NEURON MODELS
AND INTERCONNECTION PATTERNS
By

David Barnard Pierce

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Electrical Engineering

1991

__’///

\’l) 1‘

ABSTRACT

ANALYSIS OF
NEURAL NETWORK RESPONSE
WITH VARIED NEURON MODELS
AND INTERCONNECTION PATTERNS

By

David Barnard Pierce

This work addresses the performance of a neural network algorithm for constrained
optimization. The network is simulated in VHSIC Hardware Description Language
(VHDL). Comparisons are made for neurons with a step response and several levels
of graded response, as well as several methods of interconnection. The response
variables include response function, gain of the response function, and number of
interconnections. The results show that a binary response is faster than a graded
response. Additional interconnections resulted in an increase in speed of the
network. Adding a capacitance factor to the input of the neuron model reduced

oscillation and increased problem solving capability. Setting the gain and / or weights

becomes the mechanism for implementing learning algorithms.

Contents

1

INTRODUCTION
1.1 Motivation .................................
1.2 Scope ...................................

BACKGROUND

2.1 VHDL ...................................
2.1.1 Origin of VHDL ..........................
2.1.2 Structure of VHDL ........................
2.1.3 Application of VHDL .......................

2.2 Artificial Neural Networks ........................
2.2.1 Artificial Neural Network Theory ................
2.2.2 Neuron Design ..........................
2.2.3 Interconnection Effects ......................

NEURON AND NEURAL NETWORK SIMULATION

3.1 Neuron Model ...............................
3.1.1 Neuron Response Functions ...................
3.1.2 Neuron Interconnections .....................
3.1.3 Neuron VHDL Code .......................

3.2 Network Model ..............................
3.2.1 Processor .............................
3.2.2 Memory ..............................
3.2.3 Network ..............................
3.2.4 Test Bench ............................

3.3 Simulation Inputs .............................

3.4 Performance Measurement Technique ..................

3.5 Procedure for Simulation .........................

SIMULATION RESULTS
4.1 Neuron Model Results ..........................
4.2 Interconnection Pattern Results .....................

DISCUSSION
CONCLUSIONS

APPENDIX A
A.1 Neuron Model VHDL Code .......................

APPENDIX B
3.1 Network VHDL Code ...........................

iii

Ni—tt-I

quqcauswwu

H

12
12
12
16

19
19
19
20
20

22
22
23

27

32
32

43
43

0' APPENDIX C 54

0.1 Input Data Files .............. ' ............ ‘. . . 54
D . APPENDIX D , 03
DJ Simulation Outputs ...... . ..................... 63

iv

List of Figures

COGNGO‘rdeNI-d

MN NNHI—‘HHHHHHHH
huswccanqamaxuwwc

25

Hopfield-Tank Network Model ....................... 9
Sigmoid Transfer Function ......................... 10
Binary Transfer Function. ........................ 13
Low Gain Ramp Transfer Function .................... 14
High Gain Ramp Transfer Function .................... 14
Low Gain Sigmoid Transfer Function ................... 15
High Gain Sigmoid Transfer Function. ................. 15
Network Block Diagram. ......................... 18
VDHL Code for Binary Response Neuron. ............... 33
VDHL Code for Binary Response Neuron with Capacitance. ..... 34
VHDL Code for Low Gain Ramp Response Neuron ........... 35
VHDL Code for Low Gain Ramp Response Neuron with Capacitance. 36
VHDL Code for High Gain Ramp Response Neuron. ......... 37
VHDL Code for High Gain Ramp Response Neuron with Capacitance. 38
VHDL Code for Low Gain Sigmoid Response Neuron .......... 39
VHDL Code for Low Gain Sigmoid Response Neuron with Capacitance. 40
VHDL Code for High Gain Sigmoid Response Neuron. ........ 41
VHDL Code for High Gain Sigmoid Response Neuron with Capacitance. 42
VHDL Code for Neural Network. .................... 44
Two Stage Interconnected Network Input Data. ............ 55
Three Stage Interconnected Network Input Data. ........... 56
Four Stage Interconnected Network Input Data. ............ 57

Two Stage Interconnected Network with Sub-optimal Path Input Data. 59
Three Stage Interconnected Network with Sub-optimal Path Input Data. 60
Four Stage Interconnected Network with Sub-optimal Path Input Data. 61

Results of Binary Neuron Model. .................... 64
Results of Low Gain Ramp Neuron Model. ............... 68
Results of High Gain Ramp Neuron Model ................ 72
Results of Low Gain Sigmoid Neuron Model ............... 76
Results of High Gain Sigmoid Neuron Model. ............. 80
Results of Binary Response Neuron With Capacitance. ........ 84
Results of Low Gain Ramp Response Neuron With Capacitance. . . . 88
Results of High Gain Ramp Response Neuron With Capacitance. . . 92
Results of Low Gain Sigmoid Response Neuron With Capacitance. . . 96

Results of High Gain Sigmoid Response Neuron With Capacitance. . 102
Results of Three Stage Interconnection Pattern and Binary Neuron. . 106
Results of Three Stage Interconnection Pattern and Ramp Neuron. . 110
Results of Three Stage Interconnection Pattern and Sigmoid Neuron. 114
Results of Four Stage Interconnection Pattern with Binary Neuron. . 118
Results of Four Stage Interconnection Pattern with Ramp Neuron. . . 121
Results of Four Stage Interconnection Pattern with Sigmoid Neuron. . 125

42 Results of Two Stage Interconnection Pattern and Sub-optimal Path. 129
43 Results of Three Stage Interconnection Pattern and Sub-optimal Path. 133
44 Results of Four Stage Interconnection Pattern and Sub-optimal Path. 137

List of Tables

thwNI-t

Results of Neuron Model Simulations ................... 23
Results of Varied Interconnections with Binary Neurons. ....... 25
Results of Varied Interconnections with Low Gain Ramp Neurons. . . 25
Results of Varied Interconnections with Low Gain Sigmoid Neurons. . 25
Results of Varied Interconnections and Nonoptimal Path ........ 26
Results for Network with Capacitance Function ............. 29

vii

1 INTRODUCTION

1.1 Motivation

The field of neural networks, although not a new field, is currently experiencing the
focus of large amounts of research [1]. This research is exploring various methods of
implementing networks, learning algorithms, different neuron models, and applica-
tions of the networks. The promise of faster solutions to difficult problems is a major
reason for the focus upon neural networks.

Because neural networks are parallel processing structures, faster computations
are expected. In addition, neural networks have fault-tolerant structures which pro-
vide additional incentive for use [2,3]. Neural networks are presently being utilized
for many existing computation problems such as nonlinear programming [4] , control
applications [5,6], pattern recognition problems [2], and Content Addressable Mem-
ory (CAM) [9], to name a few. The parallel structure of the neural network allows
fast comparison of patterns to compute a fast, ’near optimal’ solution for a pattern,
which can often be more useful than a slowly computed, optimal solution [9]. The
neural network is also effective at the ’Traveling Salesman Problem’, an np-complete
problem that is time consuming for a single processor, sequential computing machine
[11].

Despite the amount of research in the area of neural networks, there is still a
large amount of knowledge to be gained. There is work being performed on network
structures, learning algorithms, hardware implementations, and the specific neuron
models to be utilized. One basic method of affecting the computational ability of the
neural network is to vary the specific neuron response, the basic computation block
of the network [12,13,14]. Also, the configuration of the network can have effects on
the computational ability of the network. For example, the number and location of
neuron interconnections will affect the computational ability.

Presently, there exists some work on the effect of different neuron responses and

their interconnections. Models with binary output and with graded response have
been postulated [12,13]. Work has been published to show that a graded response
will result in faster computations than a binary response [12]. However, there exists
no actual simulation or comparison data for binary as well as graded response neurons.
The effect of interconnections has also not been simulated for results. To generate an

efficient network, these areas need to be explored.

1.2 Scope

Several possibilities exist for optimizing a neural network. The number of intercon-
nections can be varied, the type of network algorithm can be varied, the method of
interconnections or data transfer can be varied, and the specific neuron implemen-
tation can be varied. These parameters are the basic building blocks of each neural
network.

As will be discussed, the type of neuron implementation can have a significant
effect on the neural network and its ability to arrive at a decision in an eficient man-
ner. This thesis addresses the performance of a neural network with changes in the
implementation of individual neuron response curves and the number of interconnec-
tions.

These changes will be simulated and the results compared through the use of
VHSIC Hardware Description Language (VHDL). VHDL allows hardware to be sim-
ulated and tested on a computer without actual implementation of the neural network
hardware. This is important as actual hardware may require a great deal of money
and time to generate results. VHDL allows simulations and results to be generated
for a minimum of cost and time [16,18].

This work will explore the effects of varied neuron responses and the effect of

varied interconnection patterns.

2 BACKGROUND
2.1 VHDL

2.1.1 Origin of VHDL

As the complexity of the engineer’s design tasks increase, the methods of generating
these designs must become more sophisticated. Without tools, such as computers
and computer programs, the design task is completely manual. Manual drafting,
development of schematics, tracking of part lists, and design analyses require large
amounts of time and effort to generate and update. Manual methods are also more
prone to errors in workmanship.

Because of the increasing complexity of the design tasks, it becomes necessary to
develop tools to handle tasks like design analyses. Tools such as computer aided draft-
ing, spreadsheets, SPICE (circuit analysis tool), and finite element analysis programs
are available to aid the engineer in his/ her design task.

For these same reasons, VHSIC Hardware Description Language (VHDL) was
conceived and generated. The Department of Defense contracted for the development
of VHDL to aid in the design of VHSIC hardware, which was a major item of defense
expenditure. Without VHDL, the design of VHSIC hardware would be difficult to
generate and to document.

VHDL will be utilized because of the amount of time and resources that would be
required to prototype a neural network. To implement the necessary neural networks
for the simulations would require many integrated circuit devices and much design
work. This, in turn, requires a large amount of financial resources: By utilizing VHDL
for simulations, the neural network can be generated and its operation verified in a
fraction of the time. Any necessary changes can be coded and verified within minutes,
which would not be possible with actual hardware. The cost to simulate the network

in VHDL is only the time spent to design and code the problem.

2.1.2 Structure of VHDL

A major advantage of VHDL is the ability to utilize varied levels of hardware ab-
stractions. The code can include exact models of gates with Boolean statements and
combine these gates into a design. The code can also use hardware concepts to sim-
ulate a design. This can be performed by coding a block which performs complex
Boolean and/ or math functions on a set of data. These types of design can also be
combined when desired.

The design is represented by a behavioral or structural model, or a combination
of both. For the designer, this allows portions of uncreated hardware to be simulated
and verified prior to spending large amounts of time implementing a specific hardware
configuration. For example, a processor unit can be modeled behaviorally by sending
certain output data when particular input data is sent to the processor. This can
be coded without Boolean gate models allowing the designer to test a function or
operation prior to generating all the internal structure necessary for operation.

Two basic elements of every system represented in VHDL are performed with an
entity statement and an architecture statement. (Words that are reserved in VHDL
are shown in boldface in this thesis). These two statements define an external and
internal view of the system. The interface of the system is defined by an entity
statement, which is the external view of the system. The behavior or structure of
the system is given in the architecture statement, which is the internal view of the
system.

The external view of the system described by the entity statement has commands
to declare the system as an entity in addition to the system’s ports. Each port
defines a signal between the system and the systems around it. Signals defined by
ports can include input, output, or bidirectional signals. Other information may also
be declared in the entity statement concerning . signals and values common to other

systems.

The internal view of the system described by the architecture statement has
commands to define what parts make up the system and how these parts define
the system’s operation. The system operation can be defined by a structural or a
behavioral model. The structural model contains models of real hardware such as
logic gates and memory devices and defines their connections. Behavioral models
describe the system operation by logic statements such as if statements and loop
commands. Both may perform the same function but are represented quite differently.
For example, a counter could be implemented with Boolean gate models connected
to operate like an actual counter. Or the Boolean logic of the counter could be
implemented with if statements without detailed gate models.

Also within the achiteeture command, components (other systems) may be de-
fined, signal values may be assigned, and other operations performed. The composted
statement places a previously defined system within the architecture being defined.
The architecture of the counter example could include gate components by the in-
stantiated component statement. The componem statement includes a port map
statement which defines the signal connections between the two systems.

The architecture of the system includes the code that defines the operation of the
system. The code could include signal assignment statements that transmit a signal
on a signal path. This can be performed as a result of a calculation, or a necesary
condition being set. Also, variables can be calculated and arrays of components
defined. These operations can be performed with loops, if statements, and other
common software functions. For example, in this work the network uses a loop
statement to calculate the summation for each neuron input. The neuron decision is
made using an if statement.

The architecture statement also includes one other important concept which is not
commonly found in software packages. The process statement defines a set of code

which will run, given certain conditions. The process can be specified with sensitivity

to certain signals. The process will begin when a transaction or even occurs on a
signal defined by the process as a sensitivity signal. The process statement also has
a wait statement which suspends operation of the process when specified conditions
occur. This capability will be utilized to update the network inputs and summations
when a transaction has occurred.

VHDL allows the use of libraries for cataloging basic design elements. Elements
such as gates, standard processors, or a hardware abstraction can be stored and
recalled for use within any entity or architecture statement. The elements stored
within a library are recalled by the use command. This command can recall another
entity, or a group of entities within a sub—library.

One more significant ability of VHDL is the ability to declare packages. A pack-
age stores the basic elements of several designs and allows for use commands to
address the package contents. A package body is the unit which contains functions
or other commands of the package. System level details such as system constants,
common functions, type declarations, and other similar information can be included
in a package for use and modification for a single system.

In this work, the package and package body will be utilized to store the summation
routine for each neuron’s input. This routine will be implemented by a function com-
mand, which returns a value when called by the software. The value to be returned

is the input summation for each neuron.

2.1.3 Application of VHDL

This work utilizes VHDL as the tool for simulating the artificial neural network and for
simulating and gathering results. VHDL is a practical, realistic approach to designing
a hardware neural network. It would be difficult and expensive to build a network
with real hardware to test the network.

The neural network utilized was originally described in [15] and will be modified

for use in this work. The modifications and rationale are described in the next section.

2.2 Artificial Neural Networks
2.2.1 Artificial Neural Network Theory

Neural network theories have been in existence since 1943 [7] The first work on neural
networks concerned the mathematical concepts necessary for the network. Later work
postulated actual hardware configurations and the mathematics supporting their use
[9]. Much work has been centered around the Hopfield-Tank network [4,11,12,13,15].

Neural networks operate by computing a solution that reduces the network’s en-
ergy function to a minimum value with certain structure conditions. These conditions
allow the network to solve the problems detailed in this report. The energy function
of the Hopfield-Tank network is

E = fizzy-swem-nl)

s' j=i

b
+2 2 Z Z: (M.+1)5V-iV(-+I)i + d(--1)i-‘V°‘V(°-"")
I t 1

+2); (35:) firms.

This energy function contains terms for the network energy and the structural
constraints of the network. This equation is minimized during the operation of a
neural network. The equation forces one neuron to be enabled in each stage and
forces the overall energy of the network to a minimum when one neuron is enabled
within each stage. The last term on the right side can be ignored for a proper choice
of gain (high gain).

The choice of weighting factors between each neuron is forced by the energy equa-
tion. The equation uses the weighting factors as part of the equation and the net-
work is therefore highly dependent upon these choices. Without the proper choice of

weighting factors, the network will not compute a valid solution.

Neural network theory generally attempts to draw parallels between biological
neural networks and artificial neural networks (non-biological). Biological neural
networks are responsible for the processing within the animal brain, such as sensory
processing, pattern recognition, and speech. Because biological neural networks are
very eficient at these tasks, it would be desirable to copy the network and utilize
the network for tasks that general purpose, sequential instruction machines perform
inefiiciently. '

The use of artificial neural networks is relatively new, due primarily to the inabil-
ity of technology to produce a hardware implementation. Artificial neural networks
require a large number of processors and interconnections, which has been beyond
the capability of circuit technolog.

The ability of manufacturers and the capabilities of equipment presently allows
small neural networks to be implemented. Massively parallel structures necessary for
large neural networks are not far away. These abilities have been progressing rapidly

over the last several years [8].

2.2.2 Neuron Design

The Hopfield-Tank network emulates the biological neural network by utilizing an
operational amplifier for each neuron. Each amplifier would have inputs, both excitory
and inhibitory, from many other neurons [9]. An example is shown in Figure 1. This
network is very large for nontrivial problems. An example would be the human brain,

which is estimated to contain 1011 neurons [10].

One of the most significant aspects of an artificial neural network is the neuron
model. The neuron has been postulated as an analog processor, such as an op-amp
[12,17] with excitory and inhibitory inputs. Most models utilize a graded response to
the inputs multiplied by their respective weights.

The neuron transfer function of a biological neuron is a monotonically increasing

 

V1 V2 V3

Figure 1: Hopfield-Tank Network Model.

and odd [-g(-u)] = [g(u)] function, or sigmoid function [3]. An example is shown in
Figure 2. The transfer function is commonly defined as

g<u>= e) we»

The analog neuron model combines inputs and weights for each connection and
sums these values. A bias term is added to the summation and then the summation
value is utilized to calculate the output of each neuron.

The neuron has also been postulated as a discrete, or digital processor [8,9,15].
This model responds with a one or zero output, based on the summation of inputs
multiplied by their respective weights. This model has also been utilized in the
analysis and simulation of neural networks.

The digital model can be implemented easier than the analog version. Simple

digital logic gates may be used in place of more complex op-amp or analog gate

 

0(x) gIXl=[-;-) [E++anh(xlo):[

 

 

Figure 2: Sigmoid Transfer Function.

models. The number of transistors necessary for a standard op-amp (Texas Instru-
ments uA741) is 22 [19]. For a standard digital gate (Texas Instruments SN5400)
the number is 4 [20]. The digital version is faster at the individual component level
due to the higher speed of digital electronics, 11 nanoseconds versus 10 microseconds
[19,20] . However, the number of gate delays in a digital model is dependent on neuron
implementation and would reduce the difference in speed.

Even with the listed advantages of the digital implementation, the network solu-
tion is not necessarily faster with digital neurons. It has been shown that an analog
network would arrive at a solution faster than the digital network [14]. This would
be true for equal time constants for each individual neuron. So it would be appro-
priate to ask, ”For what networks would the analog neuron result in a more efficient

computation? What is an ideal (or sufficient) neuron model?”

10

2.2.3 Islercounection Efl'ects

A neural network utilizes massive parallelism perform computations. This parallelism
is implemented through massive interconnection between individual neurons [2,14].
Each interconnection provides a signal which is multiplied by an appropriate weight
and summed with the remaining products. This sum determines each neuron’s output
via the neuron response curve. Inhibitory and excitory inputs can be provided to each
neuron. In some cases the inhibitory input is simply not excitory.

The effect of the number of interconnections to each neuron varies. The network
response improves with additional inputs to each neuron. This is a result of each
decision point having more information to apply toward that decision. However,
additional interconnections can have a decreasing benefit as more are added [14] .

The interconnections in typical artificial neural networks are symmetric and the
interconnections to each neuron include a full stage or stages. A nonsymmetric net-
work would connect parts of a stage or stages to a neuron. This type of network
can solve specific problems but is too problem-specific and will not be utilized in this
work. The number of stages that are interconnected to each neuron will be varied
and the results used to determine the effect of increased interconnections.

The ease of implementation of a network is dependent upon the number of in-
terconnections. Fewer interconnections reduce the complexity of the network. It is
important to ask, ”What is the optimum number of interconnections, and how many

are necessary to solve a given class of problems?”

11

3 NEURON AND NEURAL NETWORK SIM-
ULATION

3.1 Neuron Model

The neuron model utilized is similar to the model used by [15]. That model im-
plemented a two-state output for each neuron. In this work, the neuron model is

expanded to several states with different response curves.

3.1.1 Neuron Response Functions

Three types of response functions are investigated for each neuron. These include
binary, ramp, and sigmoid functions. The ramp and sigmoid results will be compared
to the binary model. This will provide a method of comparison for the different
models.

The three response functions were chosen to model the biological neuron response.
The binary function is the sigmoid function with an infinite gain. It is also the
simplest model to implement. The ramp function is used because it is close to the
actual sigmoid response of a neuron but does not have a complex math model, making
it easy to implement. The sigmoid is used because it is the actual neuron response
function. The simulations of each model will result in a comparison of the ability of
each response function to arrive at a solution when used in the neural network.

By comparing the results for each neuron, the desired response can be chosen.
The simpler models can be utilized if they provide the desired solution time, or the
more complex function may be utilized if desired. The results will show the attributes
of each of the functions.

For each response function, the gain of the curve is varied. By changing the gain
of the curve, the effects of gain assumptions are determined. These variations in the
neuron model provide coverage of factors that are important to the effectiveness of

each neuron and the network.

12

 

O for x<o
gm 9(X)=
I for x20

 

 

 

Figure 3: Binary Transfer Function.

The binary neuron response function is given in Figure 3. The two ramp neuron
response functions (for both high and low gain) are shown in Figures 4 and 5. The
two sigmoid neuron response functions (for both high and low gain) are shown in
Figures 6 and 7. These figures give the actual response function coded in VHDL
to implement the neuron. The neuron output values for each state (sum of input
multiplied by weights) is listed in Appendix A.

The gain of each of the functions is chosen to have greater or lesser resolution than
the problem set. That is, the decision of a neuron can be either the same as a binary
response or the neuron decision can be at an intermediate point on the response
function. Therefore, with a high gain function the neuron response will approximate
a binary response. The low gain functions will allow neuron outputs that are not a

one 01' 3 zero.

13

DIX) //
,1 0.0 for x<-|,5

/’ 0.25 for x<-o,5
,/ 9(x]: 0.5 for x (0.5

,’ 0.75 for x< |,5

,/ I.0 for x2|.5

 

 

 

 

 

 

 

Figure 4: Low Gain Ramp Transfer Function.

 

(0.0 for x<-o.3
0.25 for x<-C.l
0.5 for x<0.l
0.75 for x<0.3
\I.0 for x203

0(x)

 

 

 

 

Figure 5: High Gain Ramp Transfer Function.

14

 

 

90‘) /
/ 0.0 for x<-I.S

/ 0.15 for x<-0.5

/ 9(x]= 0.5 for x<O.S
r‘/ 0.85 forx<|.5

’,[/ |.0for‘ x2l.5

 

 

 

Figure 6: Low Gain Sigmoid Transfer Function.

 

0.0 for x<-0.3
0.IS for x<- 0.I
g(x)= 0.5 for x<O.I
0.85 for X<0.3
I.0 for x; 0.3

  
 

9(X)

 

Figure 7: High Gain Sigmoid Transfer Function.

15

3.1.2 Neuron Interconnections

The number of interconnections for each neuron is varied to provide a comparison of
the effect on the network. The first variations include interconnecting the stage that
the current neuron is in and the previous stage. The inputs from these two stages are
multiplied by their respective weights and summed at the neuron input. The neuron
output is then computed. For the other two variations, the previous two or three
stages and the current stage are connected to the current neuron.

The change in the number of interconnections between neurons is implemented

by changing the following:

1. Modifying the initial constants for the network within the neural package mod-
ule;

2. Modifying the memory module that holds weight factors for the neurons, and;

3. Modifying the bias term for each neuron response within the neuron modules.

The variations stated will allow visibility of the effect of interconnections in this
network.
Also, the neuron models for the increased interconnection networks are similar

but utilize a larger bias value to account for greater numbers of inputs.

3.1.3 Neuron VHDL Code

Appendix A contains the VHDL code for the neurons utilized in this work. The
code implements the neurons by calculating the sum of each weight-input product.
A series of elsif statements, which represent a lookup table, provide the output for
the appropriate state.

The summation for each neuron is peformed by a function call. This function
computes the product of weight and input and sums all products. Each neuron has

a process which is sensitive to the inputs. When a transaction occurs on an input

16

line, the neuron’s process is initiated. The process stops when the output has been

calculated until the next input transaction.

3.2 Network Model

The network model follows that of [15]. The network is setup as a processor with a
neural network as a coprocessor. The neural network receives weights from memory
as well as present neuron outputs and computes the solution. The results are then
passed to the processor. A convergence sensor is employed to sense when the network
has stabilised. The tolerance can be set within the package so it is easy to modify.

A block diagram of the network is shown in Figure 8.

Appendix B contains the VHDL code utilized to initialize the network, implement

the network, and set up a test bench.

3.2.1 Processor

The processor performs a management function for the network. The initial inputs
and initial states are passed through the processor to the neurons. Results from the
neural coprocessor are passed to the processor.

The processor module has a VHDL process which is sensitive to the outputs of
the test bench module. When the test bench module sends the neuron outputs back
to the processor, the processor’s process begins. The process assigns each neuron’s
output to the input array for each neuron. This array is used by the neuron code to
compute each neuron solution. The code utilizes if and loop statements to assign the

correct neuron outputs to each neuron’s input array.

3.2.2 Memory

A memory function is implemented to store the weights for each neuron. The weights

are drawn from memory at the beginning of the simulation. They are not used again

17

 

 

 

 

Weights

 

 

 

 

Done

 

 

 

 

 

 

Neural Network Neuron Outputs

 

 

 

New Input Array Values

 

 

 

 

Convergence
Sensor
Old
Outputs
t
Test Bench ——> Processor

 

 

 

 

 

Figure 8: Network Block Diagram.

18

 

Outputs

 

during the simulation, but could be updated if a learning algorithm was implemented.

The memory module has one signal assignment statement that sends the weight
inputs for each neuron to the neuron. These weights are arranged in an array for each
neuron. This array is multiplied by the input array for each neuron to compute the

summation for each neuron.

3.2.3 Network

The network module uses a generate statement to define the network of neurons. The

generate statement uses a loop to define an array of neurons in a single line of code.

3.2.4 That Bench

A common method for running simulations in VHDL involves the use of a test bench.
This test bench contains an entity, reference to modules in the library, and network
stimulation inputs. The test bench also provides signals for observation at the output.
The test bench used initializes the neuron states and begins the simulation.

The test bench module has a process that is sensitive to the neuron outputs. The
test bench takes the neuron outputs and sends them to the processor to be assigned

to the neuron inputs for the next cycle of calculations.

3.3 Simulation Inputs

Several data files are utilized to investigate the operation of the network. These data
files contain the weight factors utilized for each of the problems simulated. The data
(weights) are contained within the memory module. This is due to the inability of
VHDL to utilize input files with real numbers as data. These data files provide a
typical example of a problem to be solved by the neural network. The data files
contain generated cost values which simulate network constraints. The network will

seek a solution that reduces the cost along the path from start to finish.

19

The values utilized in the data sets consist of seven different values. The values,
0.0 and -9.0, either turn off or enable a neuron. This allows a starting point and
ending point to be defined for the network. The first and last stages contain these
points. The other five values are weights utilized for each neuron. These values will
enable a neuron when the weight value is less than the bias value of the neuron.

Appendix C contains the data files used for the simulations.

3.4 Performance Measurement Technique

The performance of each neuron model is measured by the response time of the
network. Each neuron model is implemented with the base network and the results
are compared. The results include the final state (if one is reached) and the number
of cycles necessary to reach that state. The final state must be a valid solution or the
compute time will be meaningless.

Because the time between each update of the network state is 10 nanoseconds (10
nanoseconds was selected to represent five gate delays in 1 to 2 micron technology),
this is the amount of time for one cycle. Using a time other than 10 nanoseconds would
provide a more exact simulation for a particular network and particular neurons. To
determine the number of cycles for convergence of the network, the setup time is
subtracted from the completion time, and this result is divided by the cycle time. In
this report the setup time is 4 nanoseconds, which is the time to feed the weights and

initial conditions to the neurons.

3.5 Procedure for Simulation

The following steps are taken for each simulation run.

Step 1: Determine Neuron Output. The appropriate response curve and the
discrete output levels are determined for each neuron.

Step 2: Code Neuron Output. The output levels are coded into the VHDL neuron

model as a lookup table.

20

Step 3: Determine Weights and Bias. The weights for each input for each neuron
are determined. The bias for the neuron response is also determined. The weights
are then coded into the VHDL memory model and neuron model.

Step 4: Edit VHDL Network File. Edit the network memory and initialization
code to utilize the appropriate neuron model, including weights and bias. The current
neuron model is added to the network code and then the file is ready to be simulated.

Step 5: Simulate VHDL File. Next, the file is put into the VHDL analyzer and
the code is compiled and used for simulation. The report output is printed as the last
part of this step.

Step 3: Analer Output. The output file is analyzed to verify that a valid solution
has been computed. If no valid solution has been calculated, return to step 3 and
repeat process. If a valid solution is computed, the number of cycles is calculated and
the final state is printed.

21

4 SIMULATION RESULTS
4.1 Neuron Model Results

The neuron model results show that the binary response function provided the fastest
network solution time. The ramp and sigmoid response functions were as fast when
the gain was set very high approximating a binary response. When the gain was set
to a lower value, the network would oscillate.

Many simulations were performed with the network and various problem sets. The
weights and bias values were varied to arrive at a network and problem set that had
a solution. The majority of the simulations resulted in oscillation or a solution set
that was not valid.

The network would arrive at invalid solutions or oscillate for two reasons. The
network was found to be very susceptible to the value of the gain of the neurons
and to the value of the neuron bias used. The neuron bias, if not set to a value
less than the weights in the nonoptimal path, results in many neurons being enabled
incorrectly (including neurons not in optimal path). If the bias is set to a value more
than some weights in the optimal path, then no solution will exist for some stages of
the network.

The oscillation of the network occurred when the gain of the graded response
neuron models was set to a low value. A high value of gain results in a response
function that closely approximates a binary function with either a 0.0 or 1.0 output.
Therefore, a low value of gain is a value that allows a resultant neuron decision to
be other than 0.0 or 1.0. For example, if the sum of products at the neuron input is
close to 0.0, then the resultant neuron decision is very close to 0.5.

This neuron then provides this output multiplied by a negative weight, to other
neurons. If more than one neuron has other than a 0.0 or 1.0 output, other neurons
are highly negatively driven. This results in a 0.0 output for many neurons in one

cycle. The following cycle then contains many neurons with a 1.0 output (because

22

Table 1: Results of Neuron Model Simulations.

 

 

 

 

 

 

 

 

 

Response Function Solution Time
Binary 15 Cycles
High Gain Ramp 15 Cycles
High Gain Sigmoid 15 Cycles
Low Gain Ramp No Solution
Low Gain Sigmoid N 0 Solution

 

there is no negative bias). The network will continue to oscillate in this way because
the low gain does not provide decision-making capability.

The high gain value provides only two decision points, so that each neuron has
good decision making capability. This results in a solution in the fastest time. The
weights and neuron bias values must be set to allow the binary neurons to arrive at
a solution.

The problem set utilized allows the binary response network to arrive at a decision
and also allows the graded response networks to arrive at a solution with the proper
bias and gain.

The results are shown in Table 1. The neuron model response and number of
cycles needed to converge are given. Appendix D contains the actual neuron outputs

with times for examination.

4.2 Interconnection Pattern Results

The network results for varied interconnection patterns showed an increase in speed
for increased numbers of interconnections. The increased numbers of interconnections

provided each neuron with a greater knowledge of the problem set. This allows the

23

first stages to compute faster. The starting point is used by the first stages as part of
the calculation and increased interconnections allowed more stages to use this known
value, thus decreasing the time for these stages to compute.

The final stage also computed faster with more than two stage interconnection.
This is due to the last stage having information from previous stages which have
already reached a solution, in addition to the previous stage.

The increased number of interconnections allowed the low gain sigmoid neuron
network to compute a solution, although this was not possible with only two stages of
interconnections. The low gain ramp neuron network oscillated as with the two stages
of interconnections. The additional knowledge combined with the better decision
making capability of the sigmoid function resulted in a solution where the ramp
function did not.

The difference in solution time between the low gain sigmoid and the binary
response networks decreased as more interconnections were added. This is due to the
added knowledge of each neuron overcoming the amount of ability of each response
function to make a decision.

Table 2 shows the simulation results for the binary neuron network. Table 3
shows results for the low gain ramp neuron network. Table 4 shows the results for
the low gain sigmoid neuron network. Appendix D contains the actual neuron outputs
with times.

A second problem was simulated with the varied number of interconnected stages.
This second problem contained a sub-optimal path between three stages that was
not part of the overall optimal path. This would show the capability of the network
to handle additional problems sets. Many applications of a neural network include
sub—optimal paths, for which fast, near optimal solutions are desired.

This additional problem set utilized a sub-optimal path between stages three,

four, and five. The sub-optimal path is therefore two stages long. The two- stage

24

Table 2: Results of Varied Interconnections with Binary Neurons.

 

 

 

 

 

 

Interconnections Solution Time
Two Stages 15 Cycles
Three Stages 13 Cycles
Four Stages 9 Cycles

 

Table 3: Results of Varied Interconnections with Low Gain Ramp Neurons.

 

 

 

 

 

Interconnections Solution Time
Two Stages No Solution
Three Stages No Solution
Four Stages No Solution

 

 

Table 4: Results of Varied Interconnections with Low Gain Sigmoid Neurons.

 

 

 

 

 

Interconnections Solution Time
Two Stages No Solution
Three Stages 19 Cycles
Four Stages 13 Cycles

 

 

25

 

 

 

Table 5: Results of Varied Interconnections and Nonoptimal Path.

 

 

 

 

 

Interconnections No Solution
Two Stages No Solution
Three Stages 13 Cycles
Four Stages 9 Cycles

 

 

 

interconnected network could not compute a solution for this problem set. The three-
and four-stage interconnected networks computed a solution in the same time as the
problem set that did not include a sub-optimal path.

The two-stage network cannot arrive at a decision in the stage that has the two
locally optimal paths. This stage sees two paths of a weight that is less than the
neuron bias, thus enabling two neurons in this stage. The network solution requires
that only one neuron is enabled per stage. So the effect of the suboptimal path is
to cause the network to oscillate at the stage where two locally optimal paths exist.
The larger two networks have enough interconnections, or information, to see a larger
solution set and resolve which of the two paths is the correct choice.

The three- and four-stage networks computed a solution in the same time as the
problem set without the sub-optimal path. This shows that a solution can be com-
puted for a network with an optimal path in a certain time and then sub-optimal
paths can be introduced with the same solution time expected. The number of inter-
connected stages necessary for the problem set is the only variable that needs to be
determined.

The results for this additional problem are shown in Table 5. The results of the

simulation are given in Appendix D for examination.

26

5 DISCUSSION

Based upon the opening text, the simulation results for the varied neuron models were
not expected. The graded response neuron models were not faster than the binary
neuron models. In some cases the graded response neuron models could not achieve
a solution.

The binary model proved to be easy to implement and simulate, whereas the
graded response models were not. The graded response models were more diflicult
to implement due to variables such as gain and the number of discrete steps. These
variables required several trials to achieve a valid solution.

The weights and bias of the neurons were revised many times during the simulation
of each different network. This was necessary to make the network converge to a valid
solution in a minimum amount of time. The network must be adjusted in this way
for each new problem, as each problem set must utilize different weight values and
different bias values. The bias values could be set at a particular value, if the weight
values were multiplied by some factor. This would put each problem set within a
particular range and allow the same bias value to used for each simulation.

The majority of the results acheived through the successive trials were oscillating
networks. The bias and / or weights would combine to drive all neurons to a high state,
then a low state, and so on. By use of a problem that has only one optimal path,
the network would converge. The optimal path must also contain each of the locally
optimal paths. Locally optimal paths not in the optimal path could only be present
if the number of interconnected stages could effectively ’step over’ this suboptimal
solution in search of the overall optimal path.

The network would not converge if the gain of the neuron response function was
low enough to allow the neurons to make ’fuzzy’ or intermediate decisions. This
caused the network to have final solutions that did not meet the validity test, that is,

not all neurons had a one or zero output. This occurred for a low gain, when decision

27

points did not fall on the ends at 1.0 or 0.0, but in between.

The neuron model first coded into VHDL does not implement all factors of the
neuron as proposed by [17] The capacitance term in the network was not imple-
mented. Additionally, the network does not accept negative input values. The input
values are either positive or zero. The models account for negative inputs with zero
input, but the capacitance function is not modeled.

Some further simulations were performed to determine if the simplified neuron
model caused the results to be other than expected. Each of the neuron response
functions was coded into a neuron model and a network that employed a form of
capacitance. Because VHDL is intended for digital hardware simulations, no capacitor
function is available. A capacitor was implemented by combining the previous input
value and the present input value. This is done for each neuron at the neuron input.
The VHDL code for this function is shown in Appendix A.

The capacitor implemented takes the past and present input and uses the following

calculation
N euronI nput = 0.6 (Present) a 0.4 (Post) .

This calculation limits a graded neuron output from going 1.0 to 0.0, and vice
versa. This allows the network to retain some memory and keep neurons from oscil-
lating in certain cases. The past value was also limited to a maximum absolute value
equal to the value of the neuron bias used. This prevents a highly negative input
from keeping the neuron off at all times.

The simulations were performed on the modified network and results taken. The
modified network results are shown in Table 6. The network did converge in one
additional case. However, the network did oscillate for the low gain ramp function
as before. The low gain sigmoid function converged to invalid solutions, but did so
relatively close to the values of the valid solution. For example, many stages with
the low gain sigmoid response had a result of 0.85, and the valid result was 1.0. The

28

Table 6: Results for Network with Capacitance Function.

 

 

 

 

 

 

 

 

 

Response Function Solution Time
Binary 15 Cycles
High Gain Ramp l6 Cycles
High Gain Sigmoid 16 Cycles
Low Gain Ramp No Solution
Low Gain Sigmoid 26 Cycles

 

result could be interpreted as a 1.0 because the other values in these stages were 0.0.
This network took longer to converge than the binary network, but did converge.

The capacitance allowed the network to retain memory and arrive at a solution.

29

6 CONCLUSIONS

The work performed showed that neural networks .and their solution time is very
strongly dependent upon the neuron model and the interconnection pattern. The
speed of the network can be increased by modifying the neuron model and increasing
the number of interconnections.

The best network in the cases studied in this work utilizes neuron models with a
response function approximating a binary neuron. This neuron model provides the
fastest solution because no intermediate cycles were needed to converge. The graded
response neuron models require extra cycles to reach the decision point because of
intermediate ’decisions’ by each neuron made during simulation.

By increasing the number of interconnections, the difference in speed between the
binary and low gain sigmoid function is decreased. This decreases the need to employ
a binary neuron. A neuron with a moderate to high gain could be employed in a
network with many interconnections with little or no penalty in speed.

The sigmoid function would also provide more information to a learning algorithm.
The learning algorithm can use this information to determine the magnitude of the
neuron input. The learning algorithm can also reset the gain within this network to
allow modest gain at the start and increase the gain as a solution is near.

Additionally, the best network contains many interconnected stages. This also
results in a faster solution time. Additional interconnected stages also eliminate
solutions that are not part of the overall optimal path because of the larger picture
seen by each neuron.

By adding a capacitance term to the input of each neuron, it was shown that
the time to reach a solution for the network was increased. However the network
would cease to oscillate and, if interpretated correctly, could provide a solution where
a simplified network would oscillate.

The hardware available today can be employed in a digital model of a neuron

30

for the best network. The slower analog components and slower speed of the graded
response networks does not compare to the binary neuron implemented with digital
gates. The benefit of the analog network would have be to great in the area of learning
algorithms to make the analog network desirable.

As research on learning algorithms grows, the optimal neuron model may be a
graded response function with the gain set to a high enough value to allow a small
degree of the switching decision to be on the response function and utilized in the
learning algorithm. It may also be advantagous for the learning algorithm to revise
the gain of the neuron model. Although this is not employed in biological neural

networks, an artificial neural network could put this increased capability to use.

31

APPENDIX A
A.1 Neuron Model VHDL Code

The response curves are modeled in VHDL by incorporating a simple lookup table.
Each table was copied to the network code and a simulation run was made. Figures

9 through 18 show the VHDL code for each neuron model.

32

-ModelofaNcuronBIcmeot

use wakNem'aIJackacull;
entity Neln'aLeIement is
Pm (
Stimulus : in My:
Weights : in InpuLAnay:
Output : out Real :- 0.0);
end NeuaLElemcot;

nchitscun'e behavitx of mm is
basin
NeuraIProccss :
MStimulus'Transactioo)
' variable Sum : Real:
basin
Sum :- CalculatcSum(Stimuhis, Weights) + 1.9;
if Sum > (0.0) also
Output <- 1.0 after 4 as;
else

quiut<-0.0atter4ns;
endif;
eodprocess;
eudbchavior:

Figure 9: VDHL Code for Binary Response Neuron.

33

-ModelofaNetnonEIemeut

use WetlaLPackagsall:
entity NeraLsIemeot is

( ,
l"Stimulus : in Input_An'ay;
Weights : in lnpuLAnay;
Output : out Real :- 0.0):
all NeraLEIcmcnt;

rcldtecuaebchsviorofnstnfl.elctneutis
componeotCapacitorponQastz'atRealzPresau:onReal):
franzapacitoruscetttttyquW).
signaIPast_Otn:ReaI:
signalPrescm_In:ReaI;

l’33.: Capcitor port um (Past_0ot. PrescutJn);
NeuraIPtosess °

MStimulus‘Trnuaction)
vaiable Sum : Real;
vniablc Cap_VaIue : Real:
variable In_Veltage: Real:

Sum :- CalcuhtcSum(Stimulus. Weights) + 1.9;
CqLVaIus :2 PresanIn;
if Cap_VaIue < -l.9 then Cap_VaIue :- -l.9:
end if;
In.Voltsge ;. (0.6‘Sum) + (0.4‘CqLVaIue);
if In.Voltage > (0.0) then
Output <- l.0 after 4 as;
else

OutputaODafteMns;
endif;
Past_0m<-Sum;

eadprocess;
eodbehavior:

Figure 10: VDHL Code for Binary Response Neuron with Capacitance.

-ModslofaNeurooEIement

use wakNeuraLPackagcall;
entity NemaLeIemeot 13

PO" (
Stimulus : in InpuLAmyi
Weights : in InpuLAnay;
Output : out Real :- 0.0);
at! WI;

architccunchehavia'otnetnaLelemattis
basin

NemalProcess: .
MStimulus’Tmisacuon)
variable Sum : Real:
sum :- CabulatsSunKStimulus. Willi") + 1'9;
if Sum < (45) the“
01mm: <- 0.0 after 4 as;
chit Sum < (-0.5) then
Output <3 0.25 3ft“ 4 03;
clsif Sum < (0.5) then
Output <- 0.50 after 4 as:
elsif Sum < (15) M
Output <- O.75 aha 4 as;

else
Ornate-1.0aflsr4n8;
audit”:

eodptocess;
endhehavior:

Figure 11: VHDL Code for Low Gain Ramp Response Neuron.

35

-ModelofaNeuronElsment

use WearaLPackageall;
unity NewaLelcmetu is
90M
Stimulus: in 1:1pr
Weights: in InpuLAnay;
quan: out Real :- 0. 0):
end Neta'aLElement;

rchitecunehehaviorofnetuaLeluneluis
componentCapacittrmesstthealzPresmtzoutReal):
endcornponent;
foraIl:Capacitoruscentitywuk.anciu(hehavior):
signalPast_Out:Real;

lmaignall’tvesentJn:Real:

gm

cm:Capacitorportm(Past_Out.Ptesent_1n);
NeuralProcesr

MStimulus‘Tmsaction)
variable Sum : Real;
variable Cap_Value : Real;
variable In_VoItage: Real:
basin
Sum. - CalculateSum(Stimuhis, Weights) + 1.9;
CqL Value :8 In:
“C71?- _Va1ue < ~1.9 then Cap_VaIue :- -1.9;
and ' :
In_VoItage :8 (0.6'Sutn) + (0.4‘CQ_VaIue);
if In_VoItage < (-l.5) then
Output <- 0.0 after 4 ns;
elsif In_Voltage < (-0.5) then
Output <- 0.25 after 4 ns;
elsif In_VoItage < (0.5) then
Output <- 0.50 after 4 as;
elsif In.VoItage < (1.5) then
Output<= 0.75 after-4113;
else Output e- 1.0 after 4 ns;

Figure 12: VHDL Code for Low Gain Ramp Response Neuron with Capacitance.

36

-ModclofaNetuonElcment

use wukNem'aLPackageall:
entity NetuaLeIemcnt is

port (
Stimulus : in InpuLArlay;
Weights: in Input_Array;
Outpm : out Real :- 0.0);
end We

rchitecuue helnvir of netnaLelement is
basin
NeuraIPtocess: '
poccss(Stimqus"1'rmsaction)
variable Sum : Real;

bean

Sum :- CalculateSum(Stimulus. Weights) + 1.9;

if Sum < (-0.3) then
Output <2 0.0 after 4 as;

elsif Sum < (-0.1) that
Outpm <- 0.25 after 4 as;

elsif Sum < (0.1) then
Olllptlt <3 0.50 after 4 as;

clsit‘ Sum < (0.3) then
Output <- 0.75 after 4 us;

else

OutputcLOafieMns;
codif;

endprocess;
endhehavior;

Figure 13: VHDL Code for High Gain Ramp Response Neuron.

37

-MorhlofaNetnonEIsmeot

use WWII:
mtity NeutaLeIement is

POM
St'nnulus:in1nput_Ana .
Weights: inInpuLAnay;
quln: outRcal ::-0.0)

mdNeraLElcmeot;

rchiucuue behavior of actual.elsment is
compaientCapscitcrponG'ast: mReaI;Present: corneal);
foran : Capacitor ascentityweI'Kancnameham);
signal Past_0ut : Real:
signal Present_ln : Real:
basin -
cap : Capsitor port mm (Past_0ut. PresentJn);
NemalProcess :
process(Stimulus’Transaction)
variable Sum : Real:
vniahle Cap_Value : Real:
variable In_VoItage: Real;

San :- CalculateSum(Stimqus. Weights) + 1.9;
Cap__Value :- PresetitJn:
NCaLValuc < -1.9 then Cap,Valuc :8 -l.9;
end if;
In_VoItage :1 (0.6‘Sum) + (0.4'Cap_Va1ue);
if In_Voltagc < (-0.3) then
Oumut <- 0.0 after 4 ns;
elsif In.VoItage < (-0.1) then
Output <- 0.25 after 4 as;
elsif In_VoItage < (0.1) then
Outpm <- 0.50 after 4 as;
elsif In_VoItage < (0.3) then
Output <= 0. 75 site 4 ns;
else Output e- 1.0 after 4 as;

Figure 14: VHDL Code for High Gain Ramp Response Neuron with Capacitance.

38

-ModslofaNeuronEIsment

use wonkNetnaljacksgerfl:
entity NeuraLelement is
POM
Stimulus : in Input_Amy;
Weights : in Input_An-ay;
Output : out Real ;. 0.0);
and NemlJElement;

actutccturc behavior of netnaLeIement is
beam
NemaIProsess:
process(8timuhis"l'rmsaction)
variable Sum : Real;
basin
Sum :2 CalculateSum(Stimulus, Weights) + 1.9;
if Sum < (-l.5) then
Output <- 0.0 after 4 as;
elsif Sum < (-O.5) then
Output <3 0.15 after 4 us;
elsit‘ Sum < (0.5) then
Output <- 0.50 after 4 ns;
elsif Sum < (1.5) then
Output <- 0.85 after 4 as;
else
Output <2 1.0 after 4 as;
end if;
end process;
end behavior:

Figure 15: VHDL Code for Low Gain Sigmoid Response Neuron.

39

-MotblofaNeumnElsmsnt

inc wakNetn'aLPackagcall;
entity NeInLelcment is
DOM
Stimulus: in Input_An'sy;
Weights: in InpuLAtray;
Output: outReal :2 0.0);
and NeraLElemcnt:

nehitecuuehelnvicsefnetuaLekmauis
componentCapacitcrpcttW:hReal:Ptescnt:omReaI);
endcompoocut;
franzmhauscmdtymcwm):
sigmlPast_Out:Rcal:
signslPtcscntankeal:

beam
eq:Capacitnrponmqi(Past_Out.Pressnt.1n):
NeuraIProcess'

MStimqus’Trauaction)
variable Sum : Real:
vriable Cap_VaIuc : Real:
vniahle In_Voltage: Real:

Sum :2 CalculatcSum(Stimqus. Weights) + 1.9;
_Value :2 PiesenLIn;
ifCap_Value < -1.9 that Cap_VaIue :2 -1.9;
end if;
ln_VoItage :2 (0.6’Sum) + (0.4‘CqLVaIuc);
if In_VoItage < (-1.5) then
Output <2 0.0 after 4 us;
chit In_VoItage < (-0.5) then
Output <= 0.15 aftu' 4 as;
elsif In_Voltage < (0.5) then
Output <2 0.50 afta 4 ns;
elsif1n_Voltsge < (1.5) then
Output <2 0.85 after 4 ns;
else Output <2 1.0 after 4 as;
end it;

Figure 16: VHDL Code for Low Gain Sigmoid Response Neuron with Capacitance.

40

-MotblofaNeuronEIement

use wakNemaLPacW-all:
entity NetuaLcIcment is

DOM
Stimulus : in hrput_An-ay;
Weights : in InpuLAnay:
Output : out Real :2 0.0);
aid NemaLEIemeut:

rchiectnre behavics of mm: is
basin
NemlProcess:
process(Stimuhis"I'rmsaction)
vaiahle Sum : Reel:
begin ,
Sum :2 CalcmstcSmMStimnlns. Weights) + 1.9:
if Sum < (-0.3) then
Output <- 0.0 after 4 ns;
clsif Sum < (-0.1) then
Output <— 0.15 after 4 as;
clsif Sum < (0.1) then
Output <2 0.50 after 4 as;
clsif Sum < (0.3) then
Output <- 0.85 after 4 as;

else
Output<21.0after4ns;
endif:

endprocess;
endhehsvior.

Figure 17: VHDL Code for High Gain Sigmoid Response Neuron.

41

Weights: inlnanAmy;
Output: ontRea! 30.0):
aldNetnLBlement;

uchiecnuebehavia'ofmmis
thflMmzhmm:wM;

comm ,
for all : Capacitor use unity MW);
signal Past_0ut: Real;

' sign! PreeenLIn : Real:
cap : Capacin pun M (PIILOut. min);
NennImess : ‘
MStimulus'Tmsaetion)

vuiable Sum : Real:
variable Cap_Value : Real;
variable BLVolnge: Real;
basin
Sun :- CalculateSmKStilnnlns. Weights) + 1.9;
CQJlalue :- Min;
if Cap_Value < -l.9 then CqLValne :- -l.9;
end if;
In_Vo!tage :- (0.6‘Sunl) + (0.4‘CQ_Value);
if In_Volnge < (-0.3) then
Outpm <- 0.0 after 4 n8;
elsif ln_Voltage < (-0.!) then
Output <= 0.15 silent as:
clsif In_Voltage < (0.1) then
Output <- 0.50 after 4 m;
elsif In_Volnge < (0.3) then
Output <a 0.85 after 4 ns;
else Output c- l.0 after 4 us;
end if;
Past_0ut <- Sum:
end mess:
end heavier;

Figure 18: VHDL Code for High Gain Sigmoid Response Neuron with Capacitance.

42

APPENDIX B
3.1 Network VHDL Code

The network is simulated by use of the VHDL code shown in Figure 19. This code

contains each of the parts of the network and the code necessary to implement them.

43

‘-Nemn!Pachge.Conninsnetwntkm-ndfnnctim8mn.

packageNem'aU’ackageis
comnnt Network_Stages:mnnl :-8;
eonsnmNemnantueonnmanmh-IZ;
eonsnntNetnmsJer_Snge:mnlnl:-6;
cunnnt‘l‘onlfietnms walnut-48; .
cannnt'l‘oleseneezRea! 30.1;
typeNeanAmyismnymau-almge l toTonlmmflleal;
typelnpuLAmyismyolmalmge l nNemmJntemonnwfReal;
type Weights_Man'ix is may (Nana's! range 1 to Total‘Jletnms) of mm
typeStimulns_Matrixismy (Nautilus 1 n'l‘onLNeunnnoflnanAI-ny;

finction CalculateSum (
Stimulus : Inpnt_Arn .
Weights = mm»
return Real;

“MM

m body NeuralJ’achge is
function CalculateSmn (
Stimulus : Input_Amy;
Weishts = InpuLAmy)
reunn Rea!
is
vuiable Sum : Rea! :- 0.0;

besm
for! in l n Neuroanteteonn loop
Sum :- Sum + Stimulusa) " Weightsa);
end loop:
reunn Sum;
end CaleuhseSun;
and NeuuLPacbge;

-ModelofaCapaciuBlemelu

nsewakNemn!_Pachge.all;
utityCapaeimris

”(Pastfiniealzl’resentzoutrealmom;
aIquncilor;

mmm' «Cg-ah is
Caphneess: .
WW)
beam

MOM
endplneess;
udhehavior;

Figure 19: VHDL Code for Neural Network.

-Netn'alPackage.ConninsnetworkparlnetusadfimetionSnm.

package NemaLPackage is
constant Network_Stages : noun! :- 8;
eomnnt Naranntaconn : natural :- 12;
constant Neurons_per_Stage : mun! :- 6:
constant Total_Nemorn : nann'al :- 48;
comrant Tolerence : Rea! :- 0.1;
type NaraLArray is my (Nauru! rage I to TotaLNeurons) ofReal;
type Input_A-nay is array (Naurra! range 1 to Neuron_lnteruonn) of Real;
type Weights_Mau-ix is array (Nana's! range 1 to ToraLNeurons) of Input_Array;
type Stimulm_Matn‘x is array (Nan! rage I to TotaLNemons) of IanLAr-ray;
Motion CalculateSum (
Stimulus : Input_Array;
Weights: Input_Array)
remrn Real;

“We:

package body NemaLPackage is
function CalculateSum (
Stimulus : hrputm
Weishts = InPuLAmy)
return Real

is
variable Stun : Rea! :- 0.0;
besin
for! in l toNemoanteteonn loop
Sum :- Sum + Stimulusa) ' Weightsfl);
ad loop:
return Sum;
end CalculateSum;
ad NemaLPacksgc;

- Model of a Capacitor Element
use workNeuraLPackageall;
entity Capacitor is

pat(Past: inreal; Presat : mural :-0.0);
ad Capacitor;
achitecture behavior of Capacitor is
besin

CapPIncess :

processmtst'Tmsactia)

helm

Present <2 Past:

endpmcess;
adbehavior;

Figure 19. (continued).

45

-‘lhismoduleisrneanttobeuedasamemorynholdtheinptuvahies

~(Weights).Acaualizedmemoryisvisualisedastheclnngerequired
-fa'adiffererunetworkwouldbeminimalthisway.

use wakNeuraLPachgeAll:
atity memory is

Don (
metnu'y.ourput : on Weights_Mau-ix);
ad memory;

architecture memory_arch of memory is
basin

M

beam

memory_output <- (

- Neuron Stage 1
(0.0. 0.0. 0.0, 0.0. 0.0. 0.0, 0.0. 0.0. 0.0. 0.0, 0.0, 0.0).
(-9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. ~9.0. -9.0. -9.0).
(-9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0).
(-9.0. -9.0. -9.0. -9.0. ~9.0. -9.0. o9.0. -9.0. -9.0. -9.0, -9.0. -9.0).
(-9.0. ~9.0. -9.0, -9.0, -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0, -9.0).
(-9.0. ~9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0. -9.0, -9.0. -9.0).
- Neuron Stage 2
(-2.5. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. -5.0. -5.0. -5.0. -5.0, -5.0).
(~35, 0.0. 0.0, 0.0. 0.0. 0.0. -5.0. 0.0. -5.0. -5.0. -5.0. -5.0).
(-0.5. 0.0, 0.0, 0.0. 0.0. 0.0. -5.0. -5.0. 0.0. -5.0. -5.0. -5.0).
(4.5. 0.0, 0.0. 0.0. 0.0. 0.0. -5.0. -5.0. -5.0. 0.0. -5.0. -5.0).
{-25, 0.0. 0.0. 0.0. 0.0. 0.0, -5.0, -5.0. -5.0. -5.0. 0.0. -5.0).
(4.5. 0.0, 0.0, 0.0, 0.0, 0.0. -5.0. -5.0, -5.0. -5.0. -5.0. 0.0),
-Neuron Stage 3
(~25. 4.5, 4.5. -3.5. -2.5. -3.5. 0.0. -5.0. -5.0. -5.0, -5.0. -5.0).
(4.5. -2.5, -3.5, 4.5, 35. ~25. -5.0. 0.0. -5.0. -5.0, -5.0. -5.0).
(~25. ~25, 4.5. 4.5, -3.5. -3.5. -5.0, -5.0. 0.0. -5.0. -5.0. -5.0).
(~25. -2.5. -1.5. 4.5. -2.5. -3.5. -5.0. -5.0. 5.0. 0.0. -5.0. -5.0).
(4.5. -4.5. -2.5. -2.5. -2.5. -2.5. -5.0. -5.0. -5.0. -5.0, 0.0. -5.0).
(-2.5. -2.5. -3.5. 4.5. -3.5. ~25. -5.0. -5.0. -5.0. -5.0. -5.0. 0.0).
- Neuron Stage 4
(-2.5. -3.5, 4.5, 4.5, -3.5. -2.5. 0.0. -5.0. -5.0. -5.0. -5.0, -5.0),
(-2.5. -3.5. -3.5. -3.5, 4.5, -2.5, -5.0, 0.0. -5.0. -5.0. -5.0. -5.0),
(-2.5. -3.5. 4.5. -2.5. -2.5. -2.5. -5.0. -5.0. 0.0. -5.0. -5.0. -5.0).
(-3.5. -3.5, -4.5, -2.5. -2.5. -2.5. -5.0. -5.0, -5.0. 0.0. -5.0. -5.0).
(4.5, -3.5. -2.5. «0.5. -2.5. 4.5. -5.0. -5.0. -5.0. -5.0. 0.0. -5.0).
(4.5, ~35, -2.5. -2.5. 4.5, -4.5. -5.0, -5.0. -5.0. -5.0. -5.0. 0.0),
- Neuron Stage 5
(-2.5. -2.5, -2.5, 4.5. -3.5, -2.5. 0.0, -5.0, -5.0. -5.0. -5.0, -5.0).
(4.5, 4.5, -2.5, 4.5. -1.5. 4.5, -5.0. 0.0. -5.0. -5.0. -5.0. -5.0).
(4.5. ~25. -2.5. -3.5. -3.5. 4.5. -5.0. -5.0. 0.0, -5.0. -5.0. o5.0).
(-2.5, 4.5. -2.5. -3.5. -3.5. 4.5. -5.0. ~5.0. -5.0. 0.0. -5.0. -S.0).
(4.5. -3.5, -2.5. -3.5. -3.5, 4.5. -5.0, -5.0. -5.0. -5.0. 0.0. -5.0).
(-4.5. -3.5. 4.5, 4.5, -3.5. -3.5. -5.0. -5.0. -5.0, -5.0. -5.0. 0.0).

Figure 19. (continued).

46

- Neuron Stage 6
(~45. ~35. ~25, ~25. ~45. ~35. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~45. ~25. ~35. 45. ~25. ~25. ~5.0, 0.0. ~5.0, ~5.0. ~5.0, ~5.0).
(~45, ~35. ~35. ~45. ~25. ~25. ~5.0, ~5.0, 0.0, ~5.0, ~5.0, ~5.0).
(~25. ~45. ~25. ~25, ~4.5. ~45. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~35. ~35. ~45. ~25. ~35. ~45. ~5.0, ~5.0, ~5.0, ~5.0. 0.0. ~5.0).
(~35. ~05, ~35. ~25, ~25. ~25. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 7
(~25. ~35. ~45. ~25. ~35. ~25. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(45. ~25. ~35. ~25. ~25. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~45, ~35. ~25, ~25. ~35. ~35. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~35. ~25. ~25. ~35. ~25. ~05. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~25. ~35. ~25. ~25. ~45. ~45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~45, ~45. ~25. ~45, ~35. ~35. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 8
(~25. ~35. ~45. ~05. ~25. ~35. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0, ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0, ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0, ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0, ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0));

wait;

end process:

ad memory_arch;

- A network of Neural Elemats
use wakNetu'aLPacltageall;
atity network is
Don (
Network_Stimulus : in Stimulus_Matrix;
Network_Weights : in Weights_Man-ix;
Network_0utput : out NeuraLAmy);
ad network:

uchitecture network_structtu'e of nemk is
component NeuraLNode
port ( .
Stimulus: in Input_Array;
Weights : in Input.Array;
Output : out Rea! :- 0.0);
end oomponat:

Figure 19. (continued).

47

-IrntaruiationofallNewaLNodesnthe
-Neural_E!ematdeaignunit

at; : Namfiode use entity workNetuaLelemathehavior);

forlinlto‘l’otaLNetn'onsgenerate
Nodes:Neural_Node

put my (Nmstimulua). NetworLWeightsG). Network_0utput(i) );
ad gaerate;
ad W

~~‘I‘hisunitisusedtosensetheeonvergenoeol’tllenetwta'k
-nasolutionwithinthetoleraeelimit

me workNem'aLPaekageall;
atity oonv_sasor is
Don (
Old_0utputs : in Neural_Array;
New_0utputs : in NeuraLArray;
Sensa_0ut : out integer);
ad eonv_senstr;

arehiecture sasor_belnvior ofconv.sasor is
basin

loop_prncess:

M01d_0utputs”l‘msaction)
variable Sum : intega' :- 0;
variable difference : real :- 0.0;
basin .

Sasor_0ut <3 1 after 0 ns;

Looplz forlin 1 to'l'onLNeurons loop
difference :- abs (Old_0utputs(l) ~ New_Otlputs(I));
if difference > Toleraee then

Sum :- Sum + 1;
end if;

ad loop loop];

if Sum =- 0 then

Sasor_0ut <8 0;
end if;
end process;
ad sensor_behavior;

Figure 19. (continued).

-Thisisthetoplevelassetnblyofallthemb-components.Differat
-signa!suegeneratedandfedndifferentparts.lnessence.itisthe
- Test-bench for the whole system

use wakNetnaljackageall:
atity ann_processor is
part (
Stimuli : in Neural_Array;
Otuputs : out NeuraLArray);
ad ann_processor:

Ichitectnre Nessa-Juneau of ann_procesaor is
component memory
pat (marory_output : om Weights_Matrix);
ad componat;
component network
pon<
Network_$timu!us : in Stimulus_Matrix:
NetworLWeights : in Weights_Matrix:
Network_0utput : out Neural_Array);
ad componat;
component conv_sasor
port (Old_0utputs : in Nem-aLArray;
New_0utputs : in NeuraLArray;
Sensor__0ut : out integer):
ad componat;
signal Old__0ut: NeuraLArray :8 (0.0. 0.0. 0.0. 0.0. 0.0. 0.0.
0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0.
0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0.
0.0. 0.0. 0.0. 0.0. 0.0, 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0.
0.0. 0.0. 0.0. 0.0. 0.0. 0.0);
signal New_0ut : NeuraLArrsy :- (1.0. 0.0. 0.0. 0.0. 0.0. 0.0.
0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0.
0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0.
0.0. 0..0 0.0 0.0. 0.0.0.0. 0.0.0.0.0.0.0.0.0.0.0.0.
0.0. 0..O 0.0. 0.0. 0.0 0.0);
W Done: integer :8 1;
signal Matrix _Weights: Weights_Matrix :- (
(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).
(0.0. 0.0. 0..0 0.0. 0..0 0.0. 0.0 0.0. 0.0 0.0. 0.0.0.0).

(0..0..00 0.0 0.0. 0.0. 0.0. 0.0. 0.0. 0..0 0.0. 0.0.0.0).
(0.0...00.00 0.0. 0..0 0..0 0.0.0.0. 0.0. 0..0 0.0. 0.0).
(0. 0. 0. 0, 0.0.0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).
(0. 0. 0. 0. 0.0.0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).
(0...0.00.00. 0.0. 0.0 0.0. 0..0 0.0. 0..0 0.0. 0.0.0.0).
(0.....0.00.00.00 0.0 0.0. 0..0 0.0. 0.0. 0.0. 0.0. 0.0).
(0.0.....00.00.00 0.0. 0.0. 0.0. 0.0. 0.0, 0..0 0.0. 0.0).
(0. 0. 0. 0. 0. 0. 0. 0.0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).
(0. 0. 0. 0. 0. 0. 0 ..0 0. 0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).
(0. 0. 0. 0. 0. 0. 0.0.0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).

Figure 19. (continued).

49

....................................

mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm. mmmmmmmmmmmm
000000000000000000000000000000000000 000000000000
00000000000&0flflflflflflflflfiflflflflQflflfi£flflflflfl(flflflflfl0000000
0000000000000000000000000000000000003000000000000
000000000Qflfiflfiflflfiflflflflflfiflflflfiflflfl£00000 flflflflflflflflflfiflfi
000000000000000000000000000000000000.000000000000
QflflflflflflflQflflfiflfinflflflflflfifififlflflaflflflflfiflflflfl &&&QQ&Q&&Q&Q
000000000000000000000000000000000000.000000000000
0000000000DDDflflflaflflflfiflflflfiflflflfiflflflflflflflmflflflfiflflflflflflfla
000000000000000000000000000000000000m000000000000
QAQQQQQQQDQflflflaflflflQflaflflflfiflflafiflflflflflflfl.QQQQQQQQQQQQ
0000000000000000000000000000000000008000000000000
Qflflfl£fiflflfififlflflflQOQQQQQQQQQQQ£QQQQQQAQ:flafiflflflflflfiflfifl
000000000000000000000000000000000000m000000000000
AQaflflflflfiflfiflflflfiflflQQQQfl0flQflQQQQQ£QQAAQmQflflflflflfiflflflflfl
000000000000000000000000000000000000m000000000000
Qfiflfl000000&0flflflflflflflflaflflflflflflflflflfiflflflflflS000000000000
000000000000000000000000000000000000.000000000000
0.0..0.0.0..0.0.0..0.00.0.0.090.0.0.0o0o00 0 00 ............

..Qmmm.mmm..mmmmm.099090000009

000000000000
anunnnuunauuuaauaanauaauaaaaaanunnau Qauunuauunan
000000000000000000000000000000000000 000000000000
aauauaannauanaunannauuunanaannnnaann unnaanauaaan

mmmm00mmmmmmwmmmmmmmmmmmmmmmmmmmmmmm.n mmmmmmmmmmmw

Figure 19. (continued).
50

(QQ QQ QQ 0.Q 0.0. 0.0. QQ QQ 0.0. QQ QQ 0..0)
....00 0.Q 0.Q 0.Q 0.Q 0.0. 0.Q 0..0 QQ 0.0.)
.0 0.Q 0.Q 0.0.0.0. 0.Q QQ 0.Q 0.0. 0.0).
...00 0.Q 0.0. QQ 0.Q 0.0. 0.0. 0.Q 0.0).
.0.Q 0.Q 0.Q 0.0. 0..0 QQ QQ QQ 0.0).
.0 0.Q QQ QQ QQ QQ QQ QQ 0.0).
.0 O.Q QQ 0.Q QQ 0.0. 0.Q 0.Q 0.0).
...00 0.0.0.0. 0.0. 0.0. QQ 0.Q 0.0).
...00 0.Q 0.Q 0.Q 0.Q QQ QQ 0.0).
0. 0.0. QQ 0.Q QQ QQ QQ 0.0).
0. 0.0. Q0. 0.0. 0.Q QQ QQ Q0).
...00 QQ QQ 0.Q QQ 0.Q 0.0).
...00 0.0. 0.0. 0.Q QQ QQ 0.0).
..0 0.Q 0.0. 0.Q 0.0. 0.0.0.0).
.0 0.0. QQ 0.Q 0.0. 0.0.0.0).
0.Q 0.0. 0.0. QQ QQ 0.0).
0...0 QQ 0..0 QQ 0.Q 0.0).
Q QQ 0.Q 0.0. 0.Q 0.0).
Q QQ 0.Q 0.Q 0..0 Q0).
0..Q 0..0 0.0. QQ Q0).
,tmmomummm,
.0. 0.0. 0.0. 0.0. 0.0).
.0 0.Q QQ 0.Q 0.0).
..Q0 0.0. 0.Q Q0).
.0..0 0.0. 0.0. 0..0)
0.MMW.0.%
.omummom
.0.mem0m
Q 0. 0. 0.0. 0.0).
0.0.0. 0.0. 0.0).
Q0.QQQQm.
Q 0. 0. 0.0. 0.0).
Q

0.

Q

0.

. ..9999999999.

boopbb

O
O C 0 C O O O O

9

pb

QOOOOOOOOOOOOOO

O O O O C C O O
O

OOOOOOOO

O
O O C O O O C C O O C O O O O O ‘

OOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOO

O
O
O

0. 0. 0.0. 0.0).
0. 0. 0.0. 0.0).
0. 0. 0.0. 0.0).
0. 0. 0.0. 0.0));

O O O O O O O O O O O O O O C C O O O

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOGOO

O O O

’58888888888888888888888888888888888
poooooooocoooooooooopppppppppppgppp

0
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.

0°COOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

0
0
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.

COOOOOOOOOOOOOOOOOOOOOOOOOOOO

..0
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.

OOOOOOOOOOOOOCOOOOCOOOOOO

0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
O.
0.
0.
0.
0.
0.

OOOOOOOOOOOOOOOOO

‘ O O O O O O .O O O O O O O C
O C O U C C O O C ‘ O O O O U

..0
0
0
0
0.
0
0.
0.
0.
0.
0.

pbbobbbb

. O O 0

Figure 19. (continued).

51

faall : networkuseatity WMJIIW);
{mall : memory useatity watmemorfimanmymh);
for all : conv_sernor use atity WMsasoLbehsvior);

Rain
man-Laud : memory port ma (Matrix_Weights):
rm.netwott : netwuk port ma (Ma'ix_Stnnnlns. Mnrix_Weights. Outputs):
sasor : conv_sasor port map (Old_0ut. New_0ut. Done):

pocess
variabletrnpl:integer:-l;
var'nbletrnp2:integu:-l;
varnbletmp3:integer:-l;
variabletmp4zinteger:-0:

begrn
waiton Stimuli’Trasaction until Done-1;

-Looptodostimu!usforeachneuron
AlLNem'ons_Loop:
forC_Netn'onsinltoTon!_Naronsloop

~~determinentrtnber'ofsngesforeachneur'onandset
unp3 :- Nemlmaconn/Netrrons_per_8tage - 2:

-loop ft: the number of stages
Each_Stage:
while unp3 > ~2 loop

-loopforeachneuroninstage

Update_Stimulus_Loop:
for C_Input in l to Netunns_per_Stage loop

- tbtermine current stage for calculation
tmpl :- ((C_Neurnns ~ l)/Neurons_per_Stage) ~ tmp3;

- ifsngeisnegativesetfm’properwnparomd
if trnpl < 1 tha
tmpl :- Netwak_Staga + tmpl:

ad if;
- settrnp2 forcorrectstimulusinputfromcurratnetuon

tme := ((trnpl ~ 1) ‘ Neurons_per_Stage) + C_Input;

Matrix_S timulus(C_Netu'ons)(trnp4 + C_Inptn) <- Stimuli(tmp2) after 2 us;

end loop Update_Stimulus_Loop;

-setfornextstagcofnarons
trnp3:-tmp3~l;

Figure 19. (continued).

52

- set for next stage of inputs
unp4 :a tmp4 + Neurons_per_Snge:
end loop Each_Stage:
:- 0;
ad loop A!l_Nenons_Loop;

- sad new values to convagence sensor
New_to_0!d_1.ocp:ftu i in l b Toflficurom loop
Old_0ut(i) c- New_0ut(i);
New Out(i) <2 Stimuli(i);
ad loop New_to.0ld_Loop;
ad process:
ad prooessor_structtn'e;

~-‘Iltisisthetestbenchformstingthewholesystem.WlIolesysteInis
-integratedinarm4mcesson

use wakNeuraLpackagenllz
entity est_bench is
ad test_bach:

architecture test_bench_arch of test_bach is

component ann_procasor
P011 (
Stimuli : in NeuraLArray;
Outputs : out NeuraLArray):
ad component:

srgrnlkln:Neural_Array:-
(0.0...10 0..0 0..0 0..0 0..0

QQ 0000 l.Q 1.0.0.0.
1.0. 0.0. 0.0. 0 .0. 0. Q 0 .0.
0.0. 1.0. 0.0, 0.0. 1.0. 0.0.
0.0. 1.0. 0.0. 1.0. 1.0. 0.0.
0.0. 1.0. 0.0. 0.0. 0.0. 0.0.
1.0. l.Q 0.0...1QOQ QQ

0.0. 0.0. 0..0 0.0 0.0 0.;0)
signalkOut: Neural _Array;
signalStoszooleanzsfalse;

for all : ann_processor use atity wortanmessoflmcessrnstrmnue);
basin
m:an_ruocesaorpatmqr(kln.k0ut);
9109533
been
waitonkOutuntilStop-false;
klna-kOutamr4ns;
Smpc-Tmeafter304ns:
adproceas:
adtest_bach.arch;

Figure 19. (continued).

53

APPENDIX C
0.1 Input Data Files

The input data files used for the simulations are shown in Figures 20 through 25.
These files contain the weights used for the simulations. The weights can be inter-
preted as cost values from one neuron to the next neuron. The more negative a value
is, the less likely that path will be chosen.

The values are arranged in arrays by neuron and by stage. There are six neurons
in each stage, and there are six stages. The first array of values contains the weights
for the first neuron in the first stage. The last six values in each neuron array are
weights for neurons within the same stage as the neuron being considered. The values
before these last six are from previous stages.

For example, with a three stage interconnected network, the first six values in
an array are the weights from neurons two stages prior, the next six values are from
neurons one stage prior, and the last six values are from neurons in the same stage.
The 0.0 value within the last six values is a direct feedback from the neuron to itself,
which is neither enabled or disabled because of the 0.0 value.

54

- Neuron Stage 1

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).

(~9.0. ~9.0. ~9.0. ~9.0, ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0, ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
- Neuron Stage 2

(~25. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~35. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, 0.0. ~5.0, ~5.0. ~5.0, ~5.0).
(~05, 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~45. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~25. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~45. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
-Neuron Stage 3

(~25. ~45. ~45. ~35. ~25. ~35. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~45. ~25. ~35. ~45. ~35. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~25. ~45. ~45. ~35. ~35. ~5.0, ~5.0. 0.0. ~5.0, ~5.0, ~5.0).
(~25. ~25. ~15. ~45. ~25. ~35. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~45. ~45. ~25. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~25. ~25. ~35. ~45. ~35. ~25. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 4

(~25. ~35. ~45. ~45, ~35. ~25. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~35. ~35. ~35. 45. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~35. ~45. ~25. ~25. ~25. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~35. ~35. ~45. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0, 0.0, ~5.0, ~5.0).
(~45. ~35. ~25. ~05. ~25. ~45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~45. ~35. ~25. ~25. ~45. ~45. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 5

(~25. ~25. ~25. ~45 . ~35. ~25. 0.0. ~5.0, ~5.0, ~5.0. ~5.0. ~5.0).
(~45. ~45. ~25. ~45. ~15. ~45. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~45. ~25. ~25. ~35, ~35. ~45. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~25. ~45. ~25. ~35. ~35. ~45. ~5.0, ~5.0, ~5.0. 0.0. ~5.0, ~5.0).
(~45. ~35. ~25. ~35. ~35. ~45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~45. ~35. ~45. ~45, ~35. ~35. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 6 .

(~45. ~35. ~25. ~25. ~45. ~35. 0.0. ~5.0, ~5.0. ~5.0, ~5.0, ~5.0).
(~45. ~25. ~35. ~45. ~25. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~45. ~35. ~35. ~45. ~25. ~25. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~25. ~45. ~25. ~25. 45. ~45. ~5.0, ~5.0, ~5.0, 0.0, ~5.0, ~5.0).
(~35. ~35. ~45. ~25. ~35. ~45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~35. ~05. ~35. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neurrm Stage 7

(~25. ~35. ~45. ~25. ~35. ~25. 0.0. ~5.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~45. ~25, ~35. ~25. ~25. ~25. ~5.0, 0.0. ~5.0, ~5.0. ~5.0, ~5.0).
(~45. ~35. ~25. ~25. ~35. ~35. ~5.0, ~5.0, 0.0. ~5.0. ~5.0, ~5.0).
(~35. ~25. ~25. ~35. ~25. ~05. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~25. ~35. ~25. ~25, ~45. ~45. ~5.0. ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~45. ~45. ~25. ~45. ~35. ~35. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 8

(~25. ~35. ~45. ~05. ~25. ~35. 0.0. ~5.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~9.0. ~9.0. ~9.0. ~9.0, ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
{-90, ~90. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0)):

Figure 20: Two Stage Interconnected Network Input Data.

55

- Neuron Stage 1

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. -9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0..~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
- Naron Stage 2

(0.0, 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~7.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~3.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0. 0.0. ~5.0, ~5.0. ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~9.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0. ~5.0, 0.0. ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~9.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0. ~5.0, ~5.0, 0.0).

-Neuron Stage 3

(~25. ~35. ~25. 45. ~25. ~35. ~25. 45. 4.5. ~35. ~25. ~35. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~25. 45. 45. ~35. ~35. 45. ~45. ~25. ~35. 45. ~35. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(45. ~35. 45. ~25. ~25. ~35. ~25. ~25. 45. 45. ~35. ~35. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~15. ~25. ~35. 45. ~35. 45. ~25. ~25. ~15. ~45. ~25. ~35. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(45. ~25. ~35. ~25. ~25. 45. 45. 45. ~25. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0, ~5.0, 0.0, ~5.0).
(45. 45. 45. ~35. ~25. ~25. ~25. ~25. ~35. 45. ~35. ~25. ~5.0, ~5.0. ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 4

(~25. ~35. 45. 4.5, ~25. ~35. ~25. ~35. 45. 45. ~35. ~25. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(45. ~35. ~25. 4.5. 4.5. ~25. ~25. ~35. ~35. ~35. 45. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~25. ~35. 4.5. ~25. 45. ~25. ~35. 45. ~25. ~25. ~25. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~35. 45. ~25, ~25. 4.5. ~25. ~35. ~35. 4.5. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~25. ~45. ~15. ~25. 45. ~35. 45. ~35. ~25. ~05. ~25. 45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~25. 45. ~35. ~35. ~25. 45. 45. ~35. ~25. ~25. 45. 45. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 5

(~25. ~35. ~35. ~25. 4.5. 45. ~25, ~25. ~25. 45. ~35. ~25. 0.0. ~5.0, ~5.0. ~5.0, ~5.0, ~5.0).
(~25. ~35. 45. ~15. ~25. 45. 45. 45. ~25. 4.5. ~15. 45. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~35. ~35. ~35. 45. ~25. 45. ~25. ~25. ~35. ~35. 45. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~25. ~35. 45. 4.5. 4.5. ~25. ~25. 45. ~25. ~35. ~35. 45. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~25. 45. ~25. ~25. ~25. 45. 45. ~35. ~25. ~35. ~35. 45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(45. 45. 4.5. ~35, ~25. ~35. 45. ~35. 45. ~45. ~35. ~35. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 6

(45. 45. ~35. ~25. ~25. ~35. 45. ~35. ~25. ~25. 45. ~35. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~35. ~35. ~35. 4.5. 45. 45. ~25. ~35. 45. ~25. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~35. ~25. 4.5. ~25. 45. 45. ~35. ~35. 45. ~25. ~25. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(45. 45. ~35. 4.5. ~35. ~25. ~25. 45. ~25. ~25. 45. 45. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(45. 45. ~25. ~25. ~35. ~35. ~35. ~35. 45. ~25. ~35. 45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~25. 45. ~25. ~35. ~15. 45. ~35. ~05. ~35. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 7

(~35. ~35. 45. ~25. 45. ~25. ~25. ~35. 45. ~25. ~35. ~25. 0.0. ~5.0. ~5.0, ~5.0, ~5.0, ~5.0).
(45. ~35. ~25. 4.5. ~25. ~35. 45. ~25. ~35. ~25. ~25. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0. ~5.0).
(45. 4.5. 45. ~25. ~35. 45. 45. ~35. ~25. ~25. ~35. ~35. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~35. ~15. 45. ~25. 45. ~35. ~35. ~25. ~25. ~35. ~25. ~05. ~5.0, ~5.0, ~5.0. 0.0. ~5.0, ~5.0).
(45. 45. ~25. 45. ~35. ~35. ~25. ~35. ~25. ~25. 45. 45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(45. ~25. ~25. ~25. ~35. ~25. 45. 45. ~25. 45. ~35. ~35. ~5.0. ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 8

(~25. 45, ~35. ~35. ~25. ~15. ~25. ~35. 45. ~05. ~25. ~35. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0, ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0));

Figure 21: Three Stage Interconnected Network Input Data.

56

- Neuron Stage 1

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0.

0.0. 0.0. 0.0. 0.0. 0.0. 0.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

- Neuron Stage 2

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~75. 0.0. 0.0. 0.0. 0.0. 0.0.

0.0. ~5.0, ~5.0. ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~105. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 45. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~135. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~75. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~135. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).

-Neuron Stage 3

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~25. ~35. ~25. 45. ~25. ~35. ~25. 4.5. ~45. ~35. ~25. ~35.

0.0. 5.0. ~5.0, ~5.0, ~5.0. -5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~25. 4.5. 4.5. ~35. ~35. 45. 45. ~25. ~35. ~45. ~35. ~25.
~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(0.0. 0.0, 0.0. 0.0. 0.0. 0.0. 45. ~35. 4.5. ~25. ~25. ~35. ~25. ~25. 45. 45. ~35. ~35.
~5.0, ~5.0, 0.0. ~5.0, -5.0. 5.0).

(0.0. 0.0. 0.0, 0.0. 0.0. 0.0. ~15. ~25. ~35. 45. ~35. 45. ~25. ~25. ~15. 45. ~25. ~35.
~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 45. ~25. ~35. ~25. ~25. 45. 45. 45. ~25. ~25. ~25. ~25.
~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 45. 4.5. 4.5. ~35. ~25. ~25. ~25. ~25. ~35. 45. ~35. ~25.
~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).

- Neuron Stage 4

(~25. ~35. 45. 4.5. ~25. ~35. ~25. ~35. 4.5. 4.5. ~25. ~35. ~25. ~35. 4.5. 45. ~35. ~25.
0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).

(~35. ~35. 45. 4.5, ~25. ~35. 45. ~35. ~25. 4.5. 45. ~25. ~25. ~15. ~35. ~35. 45. ~25.
~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(45. ~24. ~35. ~35. ~35. 45. ~25. ~25. ~35. 45. ~25. 45. ~25. ~35. 45. ~25. ~25. ~25.
~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(45. 45. 45. ~25. ~35. ~25. ~35. 45. ~25. ~25. 45. ~25. ~35. ~35. 4.5. ~25. ~25. ~25.
~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(~05. ~25. 45. ~25. ~35. 45. ~25. 45. ~15. ~25. 45. ~35. 45. ~35. ~25. ~05. ~25. 45.
~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(~25. ~35. ~35. ~35. 4.5. 45. ~25. 45. ~35. ~35. ~25. 45. 45. ~35. ~25. ~25. 45. 45.
~5.0, ~5.0, ~5.0, ~5.0. ~5.0, 0.0).

Figure 22: Four Stage Interconnected Network Input Data.

57

- Neuron Stage 5

(45. ~25. ~25. ~25. 4.5. ~35. ~25. ~35. ~35. ~25. 45. 45. ~25. ~25. ~25. 45. ~35. ~25.
0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0). '

(45. 45. ~15. ~35. ~25. 45. ~25. ~35. 45. ~15. ~25. 45. 45. 45. ~25. 45. ~15. 45.
~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(45. ~25. ~25. 4.5. ~25. 45. ~25. ~35. ~35. ~35. 45. ~25. 45. ~25. ~25. ~35. ~35. 45.
~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(~35. ~25. ~25. ~25. 45. ~35. ~25. ~15. 45. 45. 45. ~25. ~25. ~15. ~25. ~35. ~35. 45.
~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(~35. 45. 45. ~25. ~35. 45. ~25. 45. ~25. ~25. ~25. 45. 45. ~35. ~25. ~35. ~35. 45.
~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(~25. ~25. ~25. ~35. 45. ~35. 45. 45. 45. ~35. ~25. ~35. 45. ~35. 45. 45. ~35. ~35.
~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).

- Neuron Stage 6

(~25. 45. 45. 45. ~35. ~35. 45. 45. ~35. ~25. ~25. ~35. 45. ~35. ~25. ~25. 45. ~35.
0.0. ~5.0. ~5.0, ~5.0, ~5.0, ~5.0).

(~35. ~25. ~25. 4.5. ~25. ~35. ~25. ~35. ~35. ~35. 45. 45. 45. ~25. ~35. 45. ~25. ~25.
~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(45. ~25. ~35. ~35. ~35. ~45. ~25. ~35. ~25. 45. ~25. 45. 45. ~35. ~35. ~45. ~25. ~25.
~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(~35. ~25. 45. 4.5. ~25. ~25. 45. 45. ~35. 45. ~35. ~25. ~25. 4.5. ~25. ~25. 45. 4.5.
~5.0, ~5.0, ~5.0. 0.0. ~5.0, ~5.0).

(45. 45. ~35. ~25. ~25. ~25. 45. 45. ~25. ~25. ~35. ~35. ~35. ~35. 4.5. ~25. ~35. 45.
~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(45. ~25. ~25. ~15. 4.5. ~35. ~25. 45. ~25. ~35. ~15. 45. ~35. ~05. ~35. ~25. ~25. ~25.
~5.0. ~5.0, ~5.0, ~5.0, ~5.0. 0.0).

- Neuron Stage ‘7

(~25. ~35. 45. 4.5. ~35. 45. ~35. ~35. 4.5. ~25. 45. ~25. ~25. ~35. 45. ~25. ~35. ~25.
0.0. ~5.0. ~5.0, ~5.0. ~5.0. ~5.0).

(45. 45. 45. ~25. ~35. 45. 45. ~35. ~25. 4.5. ~25. ~35. 45. ~25. ~35. ~25. ~25. ~25.
~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(45. 45. ~25. ~25. ~35. ~25. 45. 45. 4.5. ~25. ~35. ~45. 45. ~35. ~25. ~25. ~35. ~35.
~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(~25. ~25. ~35. ~25. ~05. ~35. ~35. ~15. 4.5. ~25. 45. ~35. ~35. ~25. ~25. ~35. ~25. ~05.
~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(45. ~25. ~25. ~25. ~35. ~25. 45. 45. ~25. 4.5. ~35. ~35. ~25. ~35. ~25. ~25. 45. 45.
~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(45. 45. ~35. ~25. ~25. ~35. 45. ~25. ~25. ~25. ~35. ~25. 45. 4.5. ~25. 45. ~35. ~35.
~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).

- Netu'on Stage 8

(~25. ~05. ~35. ~25. 45. ~25. ~25. 45. ~35. ~35. ~25. ~15. ~25. ~35. 4.5. ~05. ~25. ~35.
0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.Q ~9.0. ~9.0. ~9.0. ~9.0. ~9.0));

Figure 22. (continued).

58

- Neuron Stage 1

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
- Neuron Stage 2

(~25. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.Q ~5.0, ~5.0. ~5.0).
(~35. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~15. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(45. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~25. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(45. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
-Neuron Stage 3

(~25. 45. 45. ~35. ~25. ~35. 0.0. ~5.0. ~5.0, ~5.0, ~5.Q ~5.0).
(45. ~25. ~35. 45. ~35. ~25. ~5.0, 0.0. ~5.0. ~5.0, ~5.0. ~5.0).
(~25. ~25. 45. 4.5. ~35. ~35. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~25. ~25. ~15. 4.5. ~25. ~35. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(45. 45. ~25. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~25. ~25. ~35, 4.5. ~35. ~25. ~5.0. ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 4

(~25. ~35. 45. 45. ~35. ~25. 0.0. ~5.0, ~5.0, ~5.0, ~5.0. ~5.0).
(~25. ~35. ~35. ~15. 4.5. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~35. 45. ~25. ~25. ~25. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~35. ~35. 45. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(45. ~35. ~25. ~05. ~25. 45. ~5.0, ~5.0, ~5.0. ~5.0, 0.0. ~5.0).
(45. ~35. ~25. ~25. ~45. 45. ~5.0. ~5.0. ~5.0, ~5.0. ~5.0, 0.0).
- Neuron Stage 5

(~25. ~25. ~25. 4.5. ~35. ~25. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(45. 45. ~25. 45. ~15. 45. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~45. ~25. ~25. ~35. ~35. ~45. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~25. ~15. ~25. ~35. ~35. 45. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(45. ~35. ~25. ~35. ~35. 45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(45. ~35. 4.5. 45. ~35. ~35. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 6

(~45. ~35. ~25. ~25. 4.5. ~35. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(45. ~25. ~35. 4.5. ~25. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(45. ~35. ~35. 4.5. ~25. ~25, ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~25. 45. ~25. ~25. 4.5. 45. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~35. ~35. 45. ~25. ~35. 45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~35. ~05. ~35. 4.5. ~25. ~25, ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 7

(~25. ~35. 4.5. ~25. ~35. ~25. 0.0. ~5.0, ~5.0, ~5.0, ~5.0. ~5.0).
(45. ~25. ~35. ~25. ~25. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).
(45. ~35. ~25. ~25. ~35. ~35. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~35. ~25. ~25. ~35. ~25. ~05. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~25. ~35. ~25. ~25. 4.5. 45. ~5.0, ~5.0. ~5.0, ~5.0, 0.0. ~5.0).
(45. 45. ~25. 4.5. ~35. ~35. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0. 0.0).
- Neuron Stage 8

(~25. ~35. 45. ~05. ~25. ~35. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0));

Figure 23: Two Stage Interconnected Network with Sub-optimal Path Input Data.

59

- Naron Stage 1

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
- Neuron Stage 2

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~7.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~3.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~9.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0. ~5.0. ~5.0, 0.0. ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~9.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).

-Neuron Stage 3

(~25. ~35. ~25. 4.5. ~25. ~35. ~25. 45. 4.5. ~35. ~25. ~35. 0.0. ~5.0, ~5.0, ~5.0. ~5.0, ~5.0).
(~25. 45. 45. ~35. ~35. 45. 45. ~25. ~35. 4.5. ~35. ~25. ~5.0, 0.0. ~5.0. ~5.0, ~5.0, ~5.0).
(45. ~35. 45. ~25. ~25. ~35. ~25. ~25. 4.5. 45. ~35. ~35. ~5.0, ~5.0, 0.0. ~5.0. ~5.0, ~5.0).
(~15. ~25. ~35. 4.5. ~35. 45. ~25. ~25. ~15. 45. ~25. ~35. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(45. ~25. ~35, ~25. ~25. 45. ~45. ~45. ~25. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(45. 45. 45. ~35. ~25. ~25. ~25. ~25. ~35. 4.5. ~35. ~25. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 4

(~25. ~35. 45. 4.5. ~25. ~35. ~25. ~35. 45. 45. ~35. ~25. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(45. ~35. ~25. 4.5. 4.5. ~25. ~25. ~35. ~35. ~15. 45. ~25. ~5.0, 0.0. ~5.0, ~5.0, ~5.0. ~5.0).
(~25. ~25. ~35. 4.5. ~25. 45. ~25. ~35. 4.5. ~25. ~25. ~25. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~35. 45. -25. ~25. 4.5. ~25. ~35. ~35. 4.5. ~25. ~25. ~25. ~5.0, ~5.0, ~5.0. 0.0. ~5.0. ~5.0).
(~25. 45. ~15. ~25. 45. ~35. 45. ~35. ~25. ~05. ~25. 45. ~5.0, ~5.0, ~5.0. ~5.0, 0.0. ~5.0).
(~25. 45. ~35. ~35. ~25. 45. 45. ~35. ~25. ~25. 45. 45. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 5

(~25. ~35. ~35. ~25. 4.5. 45. ~25. ~25. ~25. 4.5. ~35. ~25. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~35. 45. ~15. ~25. 45. 45. 45. ~25. 4.5. ~15. 45. ~5.0, 0.0. ~5.0, ~5.0. ~5.0, ~5.0).
(~25. ~35. ~35. ~35. 4.5. ~25. 4.5. ~25. ~25. ~35. ~35. 45. ~50. ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~25. ~15. 45. 45. 4.5. ~25. ~25. ~15. ~25. ~35. ~35. 45. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(~25. 45. ~25. ~25. ~25. 45. 45. ~35. ~25. ~35. ~35. 45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(45. 45. 45. ~35. ~25. ~35. 45. ~35. 4.5. 4.5. ~35. ~35. ~5.0, ~5.0. ~5.0, ~5.0, ~5.0. 0.0).
- Nmon Stage 6

(45. 4.5. ~35. ~25. ~25. ~35. 45. ~35. ~25. ~25. 45. ~35. 0.0. ~5.0. ~5.0, ~5.0, ~5.0, ~5.0).
(~25. ~35. ~35. ~35. 4.5. 45. 45. ~25. ~35. 4.5. ~25. ~25. ~5.0, 0.0. ~5.0, ~5.0. ~5.0, ~5.0).
(~25. ~35. ~25. 4.5. ~25. 45. 45. ~35. ~35. 4.5. ~25. ~25. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(45. 45. ~35. 45. ~35. ~25. ~25. 45. ~25. ~25. 45. 45. ~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).
(45. 45. ~25. ~25. ~35, ~35. ~35. ~35. 4.5. ~25. ~35. 45. ~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).
(~25. ~15. ~25. ~35. ~15. ~45. ~35. ~05. ~35. 45. ~25. ~25. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 7

(~35. ~35. 45. ~25. 45. ~25. ~25. ~35. 4.5. ~25. ~35. ~25. 0.0. ~5.0. ~5.0, ~5.0, ~5.0. ~5.0).
(45. ~35. ~25. 45. ~25. ~35. 45. ~25. ~35. ~25. ~25. ~25. ~5.0, 0.0. ~5.0. ~5.0, ~5.0, ~5.0).
(45. 45. 45. ~25. ~35. ~45. 45. ~35. ~25. ~25. ~35. ~35. ~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).
(~35. ~15. 45. ~25. 45. ~35. ~35. ~25. ~25. ~35. ~25. ~05. ~5.0, ~5.0. ~5.0, 0.0. ~5.0, ~5.0).
(45. 45. ~25. 4.5. ~35. ~35. ~25. ~35. ~25. ~25. 45. 45. ~5.0, ~5.0, ~5.0, ~5.0. 0.0. ~5.0).
(~45. ~25. ~25. ~25. ~35. ~25. ~45. 45. ~25. 4.5. ~35. ~35. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).
- Neuron Stage 8

(~25. 4.5. ~35. ~35. ~25. ~15. ~25. ~35. 4.5. ~05. ~25. ~35. 0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).
(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0));

Figure 24: Three Stage Interconnected Network with Sub-optimal Path Input Data.

60

- Neuron Stage 1

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0.

0.0. 0.0. 0.0. 0.0. 0.0. 0.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

- Neuron Stage 2

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~75. 0.0. 0.0. 0.0. 0.0. 0.0.

0.0. ~5.0. ~5.0, ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~105. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, 0.0. ~5.0, ~5.0,: ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0, 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 45. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~135. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~75. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~135. 0.0. 0.0. 0.0. 0.0. 0.0.

~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).

-Neuron Stage 3

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~25. ~35. ~25. 45. ~25. ~35. ~25. 4.5. ~45. ~35. ~25. ~35.

0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~25. 45. 4.5. ~35. ~35. 45. 4.5. ~25. ~35. 45. ~35. ~25.
~5.0. 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0, ~45. ~35. 45. ~25. ~25. ~35. ~25. ~25. 45. 45. ~35. ~35.
~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. ~15. ~25. ~35. 45. ~35. 45. ~25. ~25. ~15. 45. ~25. ~35.
~5.0, ~5.0, ~5.0. 0.0. ~5.0, ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 45. ~25. ~35. ~25. ~25. 45. 45. 45. ~25. ~25. ~25. ~25.
~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(0.0. 0.0. 0.0. 0.0. 0.0. 0.0. 45. 4.5. 4.5. ~35. ~25. ~25. ~25. ~25. ~35. 45. ~35. ~25.
~5.0. ~5.0, ~5.0, ~5.0, ~5.0, 0.0).

- Neuron Stage 4

(~25. ~35. 45. 4.5. ~25. ~35. ~25. ~35. 4.5. 45. ~25. ~35. ~25. ~35. 4.5. 45. ~35. ~25.
0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).

(~35. ~35. 45. 45. ~25. ~35. 45. ~35. ~25. 4.5. 45. ~25. ~25. ~35. ~35. ~15. 45. ~25.
~5.0. 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(45. ~2.4. ~35. ~35. ~35. 45. ~25. ~25. ~35. 4.5. ~25. 45. ~25. ~35. 4.5. ~25. ~25. ~25.
~5.0, ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(45. 45. 45. ~25, ~35. ~25. ~35. 45. ~25. ~25. 45. ~25. ~35. ~35. 4.5. ~25. ~25. ~25.
~5.0. ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(~05. ~25. 45. ~25. ~35. 45. ~25. 45. ~15. ~25. 45. ~35. 45. ~35. ~25. ~05. ~25. 45.
~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(~25. ~35. ~35. ~35. 4.5. 45. ~25. 45. ~35. ~35. ~25. 45. 45. ~35. ~25. ~25. 45. 45.
~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).

Figure 25: Four Stage Interconnected Network with Sub-optimal Path Input Data.

61

- Nemon Stage 5

(45. ~25. ~25. ~25. 4.5. ~35. ~25. ~35. ~35. ~25. 45. 45. ~25. ~25. ~25. 45. ~35. ~25.
0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).

(45. 45. ~15. ~35. ~25. 45. ~25. ~35. 45. ~15. ~25. 45. 45. 45. ~25. 45. ~15. 45.
~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(45. ~25. ~25. 4.5. ~25. 45. ~25. ~35. ~35. ~35. 45. ~25. 45. ~25. ~25. ~35. ~35. 45.
~5.0. ~5.0, 0.0. ~5.0, ~5.0, ~5.0).

(~35. ~25. ~25. ~25. 45. ~35. ~25. ~l5. 45.45. 45. ~25. ~25. ~l5.-2.5. ~35. ~35. 45.
~5.0,-5.0 ~5..0 0..0 ~5..0 ~5.0).

(~35. 45. 45. ~25. ~35. 45. ~25. 45. ~25.-2.5. ~25. 45.45. ~35.-2.5. ~35. ~35. 45.
~5...0.~50 ~5..0 ~5..0 0..0 ~5..0)

(~25. ~25. ~25. ~35. 4.5. ~35. 45. 45. 4.5. ~35. ~25. ~35. 45. ~35. 45. 45. ~35. ~35.
~5.0, ~5.0, ~5.0, ~5.0, ~5.0. 0.0).

- Neuron Stage 6

(~25. 45. 45. 4.5. ~35. ~35. 45. 45. ~35. ~25. ~25. ~35. 45. ~35. ~25. ~25. 45. ~35.
0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).

(~35. ~25. ~25. 4.5. ~25. ~35. ~25. ~35. ~35. ~35. 45. 45. 45. ~25. ~35. 45. ~25. ~25.
~5.0, 0.0. ~5.0. ~5.0, ~5.0, ~5.0).

(45. ~25. ~35. ~35. ~35. 45. ~25. ~35. ~25. 45. ~25. 45. 45. ~35. ~35. 45. ~25. ~25.
~5.0, ~5.0. 0.0. ~5.0, ~5.0, ~5.0).

(~35. ~25. 45. 4.5. ~25. ~25. 45. 45. ~35. 45. ~35. ~25. ~25. 4.5. ~25. ~25. 45. 45.
~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(45. 45. ~35. ~25. ~25. ~25. 45. 45. ~25. ~25. ~35. ~35. ~35. ~35. 45. ~25. ~35. 45.
~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(45. ~25. ~25. ~15. 45. ~35. ~25. ~l5. ~25. ~35. ~15. 45. ~35. ~05. ~35. 45. ~25. ~25.
~5..0 ~5.0, ~5.0, ~5..0 ~5..0 0.0).

- Neuron Stage 7

(~25. ~35, 45. 45 ~35 45. ~35. ~35. 45 ~25. 45. ~25. ~25. ~35. 45. ~25. ~35. ~25.
0.0. ~5..0 ~5..0 ~5.0, ~5..0 ~5.0).

(45. 45. 45. ~25. ~35. 45. 45. ~35. ~25. 4.5. ~25. ~35. 45. ~25. ~35. ~25. ~25. ~25.
~5.0, 0.0. ~5.0, ~5.0, ~5.0, ~5.0).

(45. 45. ~25. ~25. ~35. ~25. 45. 45. 45. ~25. ~35. 45. 45. ~35. ~25. ~25. ~35. ~35.
~5.0, 5.0. 0.0. ~5.0, ~5.0, ~5.0).

(~25. ~25. ~35. ~25. ~05. ~35. ~35. ~l5. 4.5. ~25. 45. ~35. ~35. ~25. ~25. ~35. ~25. ~05.
~5.0, ~5.0, ~5.0, 0.0. ~5.0, ~5.0).

(45. ~25. ~25. ~25. ~35. ~25. 45. 45. ~25. 4.5. ~35. ~35. ~25. ~35. ~25. ~25. 45. 45.
~5.0, ~5.0, ~5.0, ~5.0, 0.0. ~5.0).

(45. 4.5. ~35. ~25. ~25. ~35. 45. ~25. ~25, ~25. ~35. ~25. 45. 4.5. ~25. 45. ~35. ~35.
~5.0, ~5.0, ~5.0, ~5.0, ~5.0, 0.0).

- Neuron Stage 8

(~25. ~05. ~35. ~25, 45. ~25. ~25. 45. ~35. ~35. ~25. ~l5. ~25. ~35. 4.5. ~05. ~25. ~35.
0.0. ~5.0, ~5.0, ~5.0, ~5.0, ~5.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0).

(~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0.
~9.0. ~9.0. ~9.0. ~9.0. ~9.0. ~9.0));

Figure 25. (continued).

62

APPENDIX D
D.l Simulation Outputs

Figures 26 through 44 show the results .for all of the simulations performed. The
results show the final outputs of the neural coprocessor and the time of completion.
The time of completion minus 4 nanseconds and then divided by 10 nanoseconds

equals the number of cycles needed to arrive at a solution.

63

TIMEL

SIGNAL NAMES ,
I
(NS) |

I

KOUTG) KOUT(5)

QWEM o.oooooos+oo
LW 1.oooooos+oo
o.oooooos+oo QWEKX)
QW QWEKX)
0.(XXXXX)E+(X) QW
o.oooooos+oo
01!!me
o.oooooos+oo
QWEo-(X)
QWE‘o-(X)
o.oooooos+oo
o.oooooos+oo
o.oooooos+oo
QWBHX)
o.oooooos+oo
o.oooooos+oo
000000013400

3007(1) Iowa)
00000005430
Loooooos+oo
00000003400
o.ooooooe+oo
o.oooooos+oo
o.oooooora+oo
o. oooooaa+oo
o. oooooua+oo
0. 0000005400
0. 000000300
o.ooooooE+oo
aoooooos+oo
aooooooa+oo
o.oooooor=.+oo
o.oooooo£+oo
o.oooooora+oo
(1000000900

KOUTG) KOUT(6)

QW o.oooooos+oo
l.QXXXDE-tw Loooooos+oo
o.oooooos+oo o.oooooos+oo
o.oooooos+oo QWEHD
o.oooooos+oo 0.(XX)(XX)E+(X)
QWE-HX) o.oooooor-:+oo
o.oooooos+oo QWOOOEI-(D
QWE-I-(X) o.ooooooa+oo
QWE-HX) o.oooooos+oo
o.oooooos+oo QWE-HI)
QWEHX) QWEHD
o.oooooora+oo QWE-t-(X)
QWHX) QWEKX)
o.oooooos+oo QWE‘KX)
QWKX) QWEHX)
QWHX) QWEHX)
QWM QWEM

assessz~
é

§§§§E§29
é
é

SIGNAL NAMES ,
mum)

é

KOUT (10)

a;

KOUTO 1)

o.oooooos+oo omoooora+oo
Looooooswo

KOUT (12)

oooooooswo 0mm QW

i

§§§§§§ssrsssrs:.o

0. oooooos+oo
o.oooooos+oo
o.oooooos+oo
o.oooooos+oo
o.oooooos+oo
0. WEI-(X)
0. WSW
0. WEI-(X)
QWE-KX)
QCKDOOOEI-(X)
QWOOOEHD
o.oooooors+oo
QWE-t-(X)
o.oooooos+oo
QW

0 .oooooomoo
1.oooooor~:+oo
LWOEM
LWEM
mooooos+oo
LWEKX)
190000015400

IMHO-(X)
o.oooooos+oo
QWM
09000001900
09000001900
o.oooooos+oo
QWHX)
o.ooooooa+oo
QWM
QWOOOEM
QWEKX)
0.00(XXXJE+(X)
o.oooooos+oo
QQXDOOW
(10000001900
QW

1 000000300

QWE-t-(X)
QW
o.ooooooa+oo
QWEND
QW
QW
QWEAX)
o.ooooooa+oo

QWEHX)

QWE‘HX)
0. 0000001900
0 .WE-t-(X)
0 .(XXXXDEKX)
Q WEAK)
OWE-HI)

Figure 26: Results of Binary Neuron Model.

LWEHX)
o.ooooooe+oo
o.oooooos+oo
QWEW
o.oooooos+oo
QWEKX)
o.ooooooa+oo
QWEM
QWE‘t-(X)
o.oooooos+oo
o.oooooor-:+oo
o.oooooora+oo
o.oooooor-:+oo
o.oooooos+oo
0.000000154-00
QW

 

 

 

 

Figure 26. (continued).

65

mm 1. storm. NAMES I

I

ors)1 mums) warm) KOUTOS) mums) xoum'l) worms)

1

01 0000000900 0.000000900 0.000000900 0000000900 0000000900 0.000000900
41 1.000000900 1000000900 1.000000900 1000100900 1.00000900 1000000900
141 0. 000010900 0000000900 0000000900 0000001900 0000000900 0000000900
241 1000000900 1.000000900 1000000900 1010000900 1000000900 1000000900
341 0. 000000900 0000000900 0000010900 0000000900 0000000900 0000000900
441 0.000000900 0000000900 0000000900 1.000000900 0000000900 0000000900
541 0000000900 0000000900 0000000900 100000900 0000000900 0 .000000900
1541 0000000900 0000000900 0000000900 1.000000900 0000000900 0. 000000900
741 0.000000900 0000000900 0000000900 1000000900 0000000900 0. 000000900
841 0000000900 0000000900 0000000900 1.00000900 0000000900 0000000900
941 0000000900 0000000900 0000000900 1.00000900 0010000900 0000000900
1041 0000000900 0000100900 000000900 1010000900 0000000900 0000000900
1141 0000000900 0010000900 0000100900 1000000900 0. 000000900 0000000900
1241 0000000900 0000000900 0010000900 1000000900 0. 000000900 0000000900
1341 0010000900 0000000900 0010000900 1000000900 0. 000000900 0.000000900
1441 0000000900 0000000900 0000100900 1000100900 0000000900 0000000900
1541 0000000900 0000000900 000000900 1.00000900 0000000900 0000000900
mas. L SIGNAL NAMES 1
04s.“ mums) KOUT(20) Itom‘m) xovrm) trauma) x0009)

01 0000100900 0001000900 0. 000000900 0000000900 0000001900 0010000900
41 1000000900 1000000900 1.00000900 1000000900 100000900 100000900
141 0.000000900 0000010900 0. 000000900 0000000900 0000000900 0000000900
241 1000000900 1000000900 1000000900 1000001900 1000000900 1000000900
341 0000010900 0000000900 0000000900 0.000000900 0000000900 0000000900
441 1000000900 1000000900 1000000900 100000900 1001000900 1010000900
541 0.000000900 0000000900 0000000900 0010000900 0000000900 0000000900
1541 0000000900 0010000900 0000000900 0000000900 1000000900 0000000900
741 0.000000900 0000000900 0000000900 0000000900 1000000900 0000000900
:41 0000000900 0000000900 0000000900 0.000000900 1000000900 0000000900
941 0000000900 0000000900 0000100900 0.000000900 1000000900 0000000900
1041 0000000900 0000000900 0010000900 0000000900 1000000900 0000000900
1141 0010000900 0. 000000900 000000900 0000000900 1000010900 0000000900
1241 0001000900 0. 000000900 0000100900 0000000900 1000000900 0000000900
1341 0000000900 0. 000000900 000000900 0000000900 1000000900 0000000900
1441 0001000900 0. 000100900 0. 000000900 0000000900 1010010900 0010000900
1541 0000000900 0000100900 000000900 0000100900 1000000900 0010000900

:‘O

§§§§E§ssrsrssa

-3_g §E§§§§ssrssrss:4

 

 

SIGNAL NAMES !
KOUTQS) Km KOUT (27) KOUTQS) KW) KOUTOO)
0000000900 0.000000900 QW 0000000900 0000000900 0.000000900
1. WW 1.00000900 1 .W 1W 100000900 1.000000900
0. 000000900 0.000000900 0. WEI“) QW 0.000000900 QWEHX)
1 .mm LW 1000000900 100000900 LW 1.000000900
0000000900 QW 0000000900 000100900 QW QWEKX)
1.000000900 LW l.QXXXDENX) LW LW 1010010900
QWEM QW QW 000000900 0.000000900 0.000000900
LWEHX) LW l.QWE't-(X) 1010000900 1010000900 LWEM
0001000900 QW QWM 000000900 QW 0010000900
0.000000900 1.00000900 QW QWM QW 0.000000900
0.000000900 LW 000mm 000010900 QW QM“
QW 1000000900 000000900 QW QWENX) QW
0.000000900 1.000000900 000000900 QW QW 0.000000900
0. 000000900 1.000000900 000000900 QW 0.000000900 QW
0. 000000900 1000000900 0000000900 0000000900 0000000900 QWEAX)
0.000000900 1000000900 0.00000900 0000000900 0001000900 0.000000900
QW 1000000900 000000900 QW 0W QW
E SIGNAL NAMES E
KOUTG 1) KOUT(32) KOUT (33) KOUTCM) KOUTOS) [(001136)
0000000900 0000000900 Q W QW 000000900 QW
1.000000900 1000000900 1.00000900 1W 1000000900 LUMBER!)
0010010900 QW 0000000900 0.000000900 QW 0010000900
LWEM LW 1.000000900 100100900 1.000000900 1000000900
0W4!) QW 0000000900 0.00000900 0000000900 0.000000900
LWEM LW URINE-14X) 100000900 LW LWEM
QWEHD QW 0000000900 000000900 0000000900 0.000000900
1.000000900 1.000000900 LW MIME-14!) 1.000000900 1.000000900
QWEM QW 0.000000900 0000000900 QW 0. 000000900
LWEHX) LW 1000000900 1000000900 1001000900 1.000000900
0.000000900 QW 0000000900 Q 00000900 0.000000900 QWEM
0010010900 QW 000000900 QWEHD 0W 1.000000900
QWEHX) 0000000900 000000900 Q WEN!) 000000900 LWE-o-(D
QWEM 0000000900 000000900 0.000000900 QWE‘HX) 1.000000900
0.000000900 QW 0000000900 0001000900 0.000000900 LW
0.000000900 0.000000900 0000000900 0000000900 Q 010000900 1010000900
QW 0W 001111340) 0000001900 0000000900 LW

 

 

Figure 26. (continued).

66

EEEEEfissssssssstg

§E§§E§rsrssssssao

 

as

 

67

 

 

Figure 26. (continued).

SIGNAL NAMES !
XOUTG‘I) Km KOUTG9) KOUT(40) KOUT(41) KOUTGZ)
QWEM 0.000000900 0.000000900 OWE-IQ) 0000100900 QW
1.000000900 1.000000900 LWEHD 1000000900 1000100900 LW
0. W QW 0000000900 0000000900 Q W 0.000000900
1000000900 LW 1.000000900 100000900 1 .W 1000000900
OWEN!) QW 0000000900 0000000900 0000010900 QWEM
1.000000900 LWW 1000000900 100000900 LWEAX) 1010000900
OWEN!) QW 0000100900 0000000900 0010000900 QWBM
1000000900 1.000000900 1000100900 1.0000900 1.000000900 MIME-14D
0.000000900 QW MIME-14X) QW 0000000900 QQXXXDE-I-(X)
1000000900 1.000000900 1.000000900 1.000000900 LW LWEKX)
0.000000900 QW QWAX) 0000000900 0.000000900 0.000000900
1.000000900 1000000900 1000000900 1.000000900 1010010900 1.000000900
QW 0000000900 000000900 QW QWEM 0. WEI-(X)
Q 000000900 0000000900 QWKD 1.000000900 0.000000900 Q WOODS-14X)
Q 000000900 0.000000900 000010900 l.QXXXJOE-I-(X) QWEW 0000000900
0.000000900 0000000900 0000000900 LW 0000010900 0.000000900
QW 0W 000000900 LW QWEd-(X) QW
5 SIGNAL NAMES E
$001743) KOUT(45) KOUT(46) KOU'I‘(47) KOUTGS)
0000000900 W44!) QW 0W 000000900 QW
1.000000900 1000000900 kW 1W 1000001900 1.000000900
0000000900 QW 0000100900 QNOOOOEHX) QW QWEHD
1.000000900 kW 1.000000900 1.000000900 LW 1000010900 ~
QWEM QW 0010000900 QWHX) QWEM 0.000000900
LWEW LW 1.000000900 1.000000900 1.000000900 1.000000900
QWEW 0.000000900 0W 0. 00000900 QW QWEM
1.000000900 LW 1.000000900 1.000000900 1.000000900 1.000000900
0.000000900 0000010900 OWE-IQ) 0 .000000900 0.000000900 0. 000000900
1.000000900 1.000000900 1000000900 1.000000900 LWEHD 1.000000900
0000000900 QW-I-(D 0000000900 0.000000900 QWE-I-Q) 0. 000000900
1001000900 1000000900 1.000000900 LWEM LW-I-(X) 11!me
0010000900 0.000000900 0.00000900 QWE-I-(X) 0.000000900 0.000000900
1.000000900 LW 1.00000900 1000100900 1.000000900 1.000000900
0000000900 QMHX) 000000900 QWEM 0. WEI-(X) QWEM
”HIDDEN” 0000000900 0000100900 0.000000900 Q WIDE-14X) 0010000900
LW 0000000900 0000000900 QW Q «WE-14X) QW

 

£§§§§ii§§§§§rrrsrtrsz~

1P

-3
§

f§§33t3§39°

§§§§§§§§E§

-§_-_-______

:3
a

some»

QW
LW
0000000900
ISM-01
W01
ISM-0!
1500000901
7500000901
7W0!
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
1500000901
ISM-01
7W1

68

SIGNAL NAMES
KOUT (10)

 

0W
1W

0000000900

Figure 27: Results of Low Gain Ramp Neuron Model.

KOUTO l)

Kan-(12)

0.001411%“)

'l'llFL
NSI

B
a

§f§flifitf§za

S
a.

§§§§§EE§§

-3

23
a

§§§§§§§§§E§trtsstsss~o

 

KOUTOS)

0000100900
LWEM
QWEHX)
LWHX)
0.000000900

1-

KOUT(19)

QWM
LWEHX)
QWEHX)
1 .000000900
0.000000900
LWM
0.000000900
1.000000900
0.000000900
1000000900
QWEM
1.000000900
QWEM
! .000000900
0.000000900
1 000000900
QWEHD
1 .000000900
0.000000900
1.000000900
0.000000900
LWEM
0.000000900

norm“)
0000000900

KOUTOS)

Q W
1.000000900
Q WEI-(X)
”DOME-14X)
QWW
ISM-01
0.000000900
2500000901
QW‘HX)
ZW-Ol
0.000000900
2W0!

000000900

KOUT (21)

Q W
l .W
Q W40)
1.000000900
0010000900
”MINER”
0000100900
1.000000900
QWEW
1.000000900
0. 000000900
1 .000000900
Q 000000900
1000000900
QWM
l.Q)(XXXE'I-(X)
0.000000900
100000900
QWM
LW'KD
0000001900
1 00000900
0000000900

SIGNAL NAMES

SIGNAL NAMES

KOUTOG)

0W
1 .000000900
0.000000900
LWKD
0.000000900
kW
0000000900
7.5011100501

KOUTOD

QWM
1 00000900

KOUT (18)

QW
LWEHX)

QWEAX) QWEHX)

1010000900

LWE‘KX)

0010000900 0.000000900

7W0!

7501000901

0000001900 0.000000900

5000000901

2500100901

QWNX) 0.000000900 0010000900

1500000901

SHINE-01

WE-Ol

0010000900 0.(XX)(XJOE+(X) 0.000000900

7501100501

SIXXXXXJE-Ol

SHINE-01

ISM-01

ISM-01

QWM 0.000000900 0.000000900

75000001301

1W0!

W01

QQXDOOBHX) QWEKX) 0.000000900

ISM-01

5000000901

W01

QW QWND 0000000900

KOUTOZ)

QW
1W
QWKX)
LWM
0000000900
1.000000900
0000000900
1000000900
000000900
LW‘KX)
0.000000900

Figure 27. (continued).

69

KM)

KOUTOQ)

000000900 QWM

1.00000900
0.000000900
1.000000900
0. 000000900
100000900
0. WEI-(X)
LW
QW
1010000900
QW
1.000000900
0.000000900
1000010900
0.000000900
1.000000900
QW-KX)
l.QXXXDE-KX)
0001000900
LWE‘IQ)
QWE‘KX)
1000000900
0000010900

LUMBER!)
0.000000900
1.000000900
0.000000900
1 .000000900
0.000000900
1000010900
0000010900
LWEM
0.000000900
LWW
0.000000900
1.000000900
QWEHD
l.QXXXflfii-Q)
QW
1.0)0000900
0000010900
1000010900
0.000000900
1.000000900
QW

ML

6

E§sssssrrs=~

§§§§§§E§§

B
9.

0000000900
1.000000900
0010010900
1 WEAK)
Q 000000900
1.000000900
Q 000000900
LWEM
0.000000900
1.000000900
QWM
1.000000900
QWEKD
LWE'I-(X)
QWOOOEI-OO
LWOEKX)
QWEHX)
1.000000900
0.000000900
1.000000900
0.000000900
1010000900
QW

000000900

Q WEI-Q)
1 0100001900
0000000900

0. W
1 .000000900
0. 000100900
LWM
0010000900
LWEKX)
0.000000900
LWHXJ
QWM
1 .000000900
Q WEI-(I)
1 000000900
0. 000000900
1.000000900
0. 00000900
100100900
Q 0100005400
1000000900
QWAX)
LWHX)
0000000900
LWE'IN
000000900

SIGNAL NAMES
KOUTQS) KOU'I‘QG) KOU'I'(27) W) W)

0000000900
LWEM
QWW
1000000900
000000900
100000900
0.000000900
100000900
0.000000900
100000900
QWKD
1.000000900
0000000900
l.QXXXXIB-I-Q)
0000000900
1.000000900
QWM
11!!me
QW
1.000000900
QW
l.QXDOOE-I-(X)
QW

0.000000900
1000000900
QWEHX)
LWEHD
0000000900
LWOEAX)
QW
1001000900
QWEM
1.000000900
Q WEI-(X)
1.000000900
Q 000000900
100000900
0. WEI-(X)
1010010900
0. 000010900
1.000000900
0 .000000900
1.000000900
0. MBA!)
1000000900
000000900

KOUT (30)

0.000000900
1.000000900
QWE'KX)
1.000000900
QWE‘HX)
LWEKX)
0.000000900
l .(XXXXDEHX)
0.000000900
1.000000900
0.000000900
LWHD
0.000000900
l.Q)OOOOEM
0.000000900
1.000000900
QWEKX)
LWEKX)
0.000000900
1.000000900
0.000000900
LWHD
0010000900

‘P

SIGNALNAMES E

l
(NS) l KOUTGI) KOU'I'OZ) [OUT (33)

«b

and
L

§§§§§§E§§E§ffiirtrfi

B
a

QWEHX)
LWEHX)
0.000000900
1.000000900
0.000000900
1.000000900
0.000000900
1.000000900
QWEM
1.000000900
0.000000900
1.000000900
QWEHX)
l.QXDOOE-I-(X)
QWE-t-Q)
1.010000900
QWEA!)
1.000000900
QWEM
1.000000900
0010000900
1.000000900
QWEMX)

0000000900
1000010900
Q 000000900

100000900

QWEHD

1.000000900

0.000000900

LMHD

QW
1.000000900
QWEW

l.Q!!WEi-Q)
0.000000900
1010000900
0.000000900
1010000900
0.000000900
1.000000900

0010000900

1010000900

0000010900

1000010900

0000100900

QWAD
LW

0.000000900
1010000900

0.000000900

1000010900

0.000000900

1.000000900

0.000000900
1.000000900
Q 010000900

100010900
QWW
100010900
QWEM
LWM
0000000900
1 .000000900

0.000000900

1.000000900

0.000000900

1.000000900

000000900

Figure 27. (continued).

70

0. OWEN!)

1.000000900
1 010000900

QWEM

LWEHX)
QWEd-Q)
l .WEAX)
0. 000000900
1.000000900
0.000000900
1.000000900

QQXJOOOE-I-(X)

1.000000900

QWOOOE-I-Q)

l.QXXXDB-Q)

QW

TIMBL

SIGNAL 14.411153 5
1
043} l

mum?) K0011”) KOU'l‘t40) KOUT(41)
0000000900
100000900
0000000900
1000000900
0000000900
1.010000900
0010000900
1001000900
0000000900
1010000900
0000000900
1010000900
0000000900
1000000900
0000000900
1000000900
0000000900
1000000900
0000000900
1000000900
0000000900
1000000900

KOUT(42)

QWEHX)
LWBW
QWEHX)
1.000000900
Q 000000900
1 010000900
0.000000900
1.000000900
0. WEI-(X)
1.000000900
Q WEI-(X)
1.000000900
0.000000900
1.000000900

QW
1010000900
0.000000900
1000000900
0.000000900
1000010900
0000100900
1 W400
Q 000000900
1 .000000900
0. 010000900
LWKX)
0.00000900
100000900
0.00000900
1.000000900
QWEKX)
1 .000000900
QWAX)
1010000900
0.000000900
1.000000900
000010900

0000000900
LW
0.000000900
1.000000900
000000900
1.000000900
0000000900

QW
1.000000900
QWEM
1.000000900
QWE'l-Q)
LWEI-(X)
QWE-t-(X)
1.000000900
QWEKX)
1111me
0.000000900
1.000000900
QW
LWEKX)
QWEHD
l .WEW
0. WEAK)
1.000000900
Q WEI-(I)
1000100900
0000100900
LWHX)

gas:

1 .WOE-KX)
0. 000000900
1.000000900
000000900
1.00me
000000900
1.000000900
0.000000900

§§§§§§E§§E§f§flfitrfi

B
a

rsrrusaazsfffiirtrfizso

gflfluflflflfl—uu
------

8
a.

KOUT(43)

0.000000900
LWEHX)
QWEHX)
LWEW
QWEHD
”MINE-(X)
QWEA'D
1.000000900
QWEM
1.000000900
QQXXXDEHX)
1.000000900
0.000000900
1010000900
0.000000900
1.000000900
QQXIJOOE-I-(X)
1.000000900
QWE'IN
1.010000900
0.000000900
1.000000900
0.000000900

KOU'IIM)

0000000900
1010000900
QW
LW
QWHD
LW
0.000000900
1 000000900
OHM-14!)
1010000900
0.000000900
1000100900
0000000900
1.000000900
0.000000900
1.000000900
0.000000900
LWEM
0000100900
LWEW
0000000900
1000000900
0.010000900

rooms)

QW
LW
QWE‘KX)
1.000000900
0.000000900
1.000000900
QWEM
1.000000900
QWEW
LOWE-(X)
0010000900
LW-I-Q)
0.000000900
l.QXJOOOE-I-(X)
QWKX)
1.000000900
0.000000900
100000900
Q WEI-(X)
1 .000000900
0. W44!)
1.00000900
000000900

SIGNAL NAMES

KOUT(46)

Q WEI-(X)
1W
0. 00000900
l.Q)OOOtE4-(X)
QWOEI-OO
1.000000900
0.000000900
1.000000900
QOQJOIXEM
100000900
0000000900

Figure 27. (continued).

71

KOUT(47)

0000001900

QW-I-Q)

0.000000900

KOUT (48)

0000000900
LW
0.00000900
1.000000900
0.000000900
l.QXXXXEHX)
0000000900
1.000000900
0.000000900
1000010900
QWEW
1.000000900
QWEW
1000000900
0.000000900
l.QXXXJOEHX)
000000054X)
1.000000900
QQX)000E+(X)
1 .WE-I-Q)
Q 000000900
1.000000900
QW

 

 

 

 

SIGNAL NAMES I

I

(NSI) I KOUTO) KOUTQ) KOUT (3) KOU'l'(4) KOUTG) KOUT(6)

0! 0000000900 QW QWEM 0000010900 QW Q 010010900

4! 1000000900 LW LW 1000100900 1.000000900 LWEKX)
14! 1000100900 QW-I-(X! QW QW QWEM 0. 000000900
24! 1.000000900 0.000000900 QW-HX) 0.000000900 0.000000900 0010000900
34! LWOEHD 0.00000900 0.000000900 0000100900 QWE-I-(X! 0. W
44! l.QXXJOOE'I-(X! 0.000000900 QW 0000000900 0.000000900 0. 000000900
54! l.QXXDOE-I-(X! 000010900 QWE-I-(X! QW 0000100900 0. 000100900
64! 1.000000900 QWHX! 0.000000900 QWE-I-(X) 0010000900 0.000000900
74! LIXXDOOE'KX) QW‘KX! 0.000000900 QW 0010000900 0.000000900
84! 1.000000900 000100900 QW 01!!me 0000000900 0010000900
94! 1.000000900 0000100900 0.000000900 QW 0.000000900 Q WEI-(I!
104! 1.000000900 0111110054!) 0.000000900 0.000000900 0.000000900 Q WEI-(X!
114! 1000000900 QW QWMX) 0001000900 QWAD 0.000000900
124! 1.000000900 QW 001000900 QWEHX) QWM QW
134! 1.000000900 QW 0000000900 DUMB-14!) 000000900 QW
144! 1.000000900 QWE-I-(X! QWM 0000000900 000010900 QWEW
154! 1.000000900 QW 000000900 0W 000000900 QW
m I SIGNAL NAMES I

I
(NS) I KOU'I'G) KOU'KS) KOUTG) KOUTOO) KOU'I'OI) KOUTOZ)

I

0! 0.000000900 0.000000900 Q W 0W 000000900 Q W
4! l.QXDOOEvI-(D 1000000900 LW 1010000900 1.00000900 LW
14! 0.000000900 QW 0 .000000900 QWM QW 0 000010900
24! 0.000000900 0. W 1000010900 0010000900 0.000000900 QWE-I-(X)
34! OWEN!) Q WEI-(X! 1.000000900 QW-I-(X! 0000000900 QIXXXXDEt-(X)
44! QWEM QW 1000000900 0.000000900 0. “0008+“! 0.000000900
54! 0.000000900 0.000000900 1000100900 0.000000900 Q WEI-(X) 0.000000900
64! QWEM QW LWM 0.000000900 0010000900 0.000000900
74! 0.000000900 QW LWEM 0000000900 0.000000900 QWEM
84! 0000010900 0. W 1.“!!!an 0.00000900 QW 0.000000900
94! 0000000900 0. WEI-Q) 1.000000900 Q 010000900 Q 000000900 QWW
104! 0000000900 QWM 1000000900 QW 0.000000900 QWOOOB-I-(D
114! 0. MOE-14X! 0.000000900 1000000900 QW Q WEI-(X) 0.000000900
124! 0. 000000900 QWEW LWM 0000000900 MIME-14!) 0.000000900
134! Q MOE-14X! 0.000000900 LWW OMXDOOEHX) 0.000000900 0.000000900
144! QWEHX! 0000010900 1010000900 0000000900 0010010900 0010000900
154! QW 0W 100000900 QW 0.00000900 0.000000900

Figure 28: Results of High Gain Ramp Neuron Model.

72

 

TIME 9 SIGNAL NAMES I
I

(NS) I KOU'I'(13) KOUIU4) KOU'I‘ (15) KOUTOG) KOUTO'I) KOUT (18)

 

 

0000000900 0000000900 0.000000900 0W 000000900 QW
4 I 1W 1000010900 1010000900 1W 100000900 1.000000900
14 I 0.000000900 QW QWEI-IX) QW QW 0.000000900
24 I 1.000000900 1.000000900 LWE'IQ) 1000000900 MINDS-IQ) 1.000000900
34 I QWEM QW 0.000000900 QW QW 0.000000900
44 I OWEN!) QW 0.000000900 LW QW 0000000900
54 I 0000000900 QW 0000000900 LWM QW 0010000900
64 I OWEN QW 0000000900 1.000000900 QW 0010000900
74 I OWE-(X) QW QW LW QW QWEKX)
84 I 0.000000900 QW QWEHX! 100000900 QW OWEN!)
94 I 0.000000900 QW QWM 1.00000900 QW 0.000000900
104 I 0000000900 0000000900 000000900 100000900 QWM QW
114 I QW 0.000000900 0W 1.000000900 QWM 0.000000900
124 ! QW 0W 0.000000900 LW Q 000000900 0.000000900
134 I QW 0.000000900 0.00000900 LW Q WEI-(X) QWE'I-(X)
144 I 0.000000900 0000000900 000000900 LW Q 000000900 0.000000900
154 I W W 0.00000900 LW 0W QW
TIME I SIGNAL NAMES I
!
“‘5?

1 zooms) 11001120) KOUT(21) x0002) KOU'I‘(23) Korma)

0 I 0000000900 000100900 QWW 0000000900 0W4“! QW
4 I 1001000900 1W LW 1W LW 1.000000900
14 I 0000010900 QW 0.000000900 0000000900 QWEd-(X! QWEM
24 I 1000000900 1.000000900 1.000000900 1000100900 1.000000900 1.000000900
34 I 0. 000000900 QW QM“ 0000000900 0000000900 Q WEI-(X)
44 I 1 .000000900 LW LWEM 100000900 1.000000900 1 WEI-IX)
54 I OWE-(X) QW QWEM 0. 000000900 QW Q 000000900
64 I OWEN!) QW QW 0. W 1010000900 0 .000000900
74 I 0000000900 QW QWM 0. 000000900 1.00000900 Q WEI-(XI
84 I OWE-(X) QW QWW 01010000900 LW 0.000000900
94 I 0001000900 QW 0000100900 000000900 1.000000900 0.000000900
104 I Q WEI-(X! 0000000900 0000000900 QW 1W QW
114 I Q MOB-14X) 000000900 000000900 QW 1010010900 QW
124 I Q 010000900 0000000900 QWEM QW 100000900 0.000000900
134 I QWEMX) 0000000900 0.000000900 QW 1001000900 QWEHX)
144 I 0000000900 0000000900 0000000900 QW 1.00000900 0000010900
154 I 0W 0W 000000900 0000000900 1W 0.000000900

Figure 28. (continued).

73

TIME;

§§§§E§ssssssrss~o

0.4
.1

reassess:

§E§§E§

 

 

 

 

Figure 28. (continued).

74

SIGNAL NAMES I
I
(NS) I KOUTOS) KOUTOG) KOUT(27) KOU'I‘(28) KW) KOU'I‘ (30)
!
0000000900 QWEHX) QWKD 0W 0001001900 QW
1.000000900 1.000000900 LW 1.000000900 1M4“) 1010000900
QOOIXXXIE'I-(X) QWBAD 0.000000900 QW-I-Q) QIXDOOOE'HX) QWEI-OO
1MB“) l.QXXXIOE-I-Q) 1000000900 LWHX) LW LWE‘KX)
QWOIDEKX) 0000010900 0000000900 00000005400 0000010900 QWE'I-IX)
1.000000900 LW 1000010900 1000001900 1.000000900 l.QXXXDEKX)
0000000900 QWEHXI QWEI-IX) 001000900 0.000000900 QWEHXI
IWE‘KX) LW 1000000900 LW‘IN LW 1.000000900
0000000900 QWEHX) 0.000000900 QWAXI 0.000000900 0.000000900
OWE-(I) 1.000000900 0W4“) 0000000900 QWEAX) QWM
0.000000900 LIXXXIOOEHX) 0.000000900 001000900 QWHX) QWEKX)
QWEI-W 1000000900 0000000900 QWEM QW-t-(X) QWEHX)
QWEW l.QXXXXIE'I-Q) 0.00000900 0000000900 0.000000900 QWEW
0.000000900 LOWE-14X! QWKX! 0.000000900 QIanEt-(X) 0.000000900
0001000900 1.000000900 0000000900 0.000000900 0.000000900 QWEI-Q)
0000000900 1000000900 000000900 0000000900 0000000900 QWEHD
QW 1000000900 000000900 0000000900 0.000000900 QWE'I-Q)
I SIGNAL NAMES I
KOU'I'G 1) KOU'I‘GZ) KOU'I' (33) KOUT(34) KOU'I‘OS) KOU'I' (36)
0.000000900 0000000900 QW 0W 0000100900 QW
LWB-t-(X) 1000000900 LW 1W 100100900 l.QXXXXIE-I-(X)
QWEHX! 0.000000900 QWEM QWAX) 0.000000900 QWE‘KX)
1.000000900 LW ' 1000000900 1000000900 LWEHD IMHO“)
QWEKX) QW 0000010900 0010000900 0.000000900 0.000000900
”DOME-(X) l.QDOME‘I-Q) 1000010900 1000000900 l.QXXIOOE-I-Q) 1.000000900
QIXIOOIDE+00 0001000900 0.000000900 0010000900 0.000000900 0.000000900
1000000900 1.000000900 LWBM 100000900 LWEKX) 1.000000900
QWEAI! QW QW 0010000900 QW 0.000000900
l.Q!)OIDE-I-IX) 1.000000900 1.000000900 1.000000900 ”HINGE-(X) 1.000000900
QWEM 0010000900 QWEKXI 0.00000900 0.000000900 0.000000900
0.000000900 QWE'I-(XI 0.000000900 QW 0010010900 1.000000900
0.IXX)OOOE+(X) QWEKX) QWKX) 0.000000900 0.000000900 LWEHD
0.000000900 QWEM 0010000900 QWEM 0.000000900 1.000000900
QOQIOOOE-t-(X) 0.000000900 QQXXXXE-I-(X) QOOIXIOOE'I-(X) 0.000000900 1.000000900
QWEKXI 0010010900 0000000900 QW 0010000900 l.QXXXIOEI-Q)
0.000000900 0000000900 0000000900 QW 000100900 1.000000900

MI-

 

 

 

 

Figure 28. (continued).

75

SIGNAL NAMES I
I
(NS) I K001 (37) KOU'I'GS) K0010” K001‘(40) [(001141) K001 (42)
I
0 I W44!) 0000000900 Q W 0000000900 000000900 0.000000900
4 I 1.000000900 1.000000900 1 .000000900 100000900 LWEM l.QXXXIOEHX)
14 I QM“ QW 0. 000000900 0010000900 0001000900 QWEAX)
24 I 1.000000900 LW 1.000000900 1.00000900 LWEM LWEM
34 I OWE-(X) QW 0.000000900 0.00000900 0.000000900 0.000000900
44 I 1000000900 LWHX) 1010000900 1.000000900 LW l.QXXXDE-I-IX)
54 I QWEM QW 0.000000900 0.00000900 0. WAX) QWEM
64 I 1.000000900 LW LWBM 1.000000900 1.000000900 LWEHD
74 I 0000000900 QW 0.000000900 Q W44!) 0. WEI-(X) QWEHX)
84 I LWEHXI LW LWEKX) LWW 1.000000900 LWEW
94 I 0000000900 QW 0000000900 Q 00000900 QWAX) QWBW
104 I 1.1!!!)me 1.000000900 LWKII 1.000000900 1.000000900 l.QXXXXIENX!
114 I 0.000000900 0010000900 000000900 0.000000900 0000010900 QWE-I-(D
124 I 0.000000900 0.000000900 QWHX) LW 0.(XXXXX)E+(X) QW
134 I 0.000000900 QWAXI 0 .00000900 1.000000900 QWEM 0010000900
144 I QWEM 0000000900 Q 000000900 1.000000900 OWEN!) QW
154 I 0000000900 0000000900 000000900 LW QW QW
11MB I SIGNAL NAMES I
I
(NS) I KOUT(43) KOUKM) K001" (45) K00'I‘(46) K00'I'(47) X001 (48)
I
0 I 0. MOE-14X! 0000010900 QW Q W Q 0000001900 0000000900
4 I 1000100900 LWEM l.QWEI-(D 1000000900 1 .000000900 1.000000900
14 I 0. WEI“) QMXJOOEHD 0.000000900 0000000900 Q 010000900 QWEKX)
24 I 1.000000900 1.000000900 1.000000900 1.00000900 1.000000900 1.000000900
34 I QWEKX! QWEW QWEI-Q) 0.00000900 0.000000900 0.000000900
44 I 1.000000900 1.000000900 1.000000900 1.000000900 1001000900 1.000000900
54 I 0.000000900 QWEAD 0.000000900 0000100900 0.000000900 QWEHX)
64 I LWEM LWEHD 1010000900 1.00000900 1.000000900 1000010900
‘74 I QWOOOEKX) 0.000000900 0.000000900 0.000000900 QW-I-IXI 0.000000900
84 I 1.000000900 LW l.QXXDOE-I-Q) 1.000000900 1.01MB“) 1000010900
94 I 0010000900 0.000000900 QWAX) 0000000900 0.000000900 QWEM
104 I 1.000000900 1.000000900 1000000900 1.000000900 LIXXXXDEHX) LWEM
114 I 0.000000900 0.000000900 0.0000900 QIXXXIOOEKX) 0.000000900 0.000000900
124 I LW LWIKX) LWAX) LW 1000000900 LW
134 I 031300244!) 0.000000900 00000900 QW 0. 000000900 0. 000000900
144 I LWEW 0000000900 000000900 0.000000900 0. 000000900 Q 000000900
154 I 1000100900 0000100900 0W4“) QW Q 000000900 Q W

§ssasssrszuo

S
a

§§§§§§EEE

I‘.’
a

4

-3

inc

iifiifitfgi’fitgggrtrg

gunned—Hummus
P
O

N
an
‘

I SIGNAL NAMES J.

I K0010) K0010) K0016) K00'1‘(4) K001'(5) K0016)

I 000010900 QW QW 0000000900 0000100900 QW
1.000000900 LWHD LW 1.000000900 1000100900 1.000000900
1.000000900 0.000000900 QW QW 0000000900 0.000000900
1.000000900 0.000000900 000000900 QW 0.000000900 QWEHD
l.QXXXXJE-t-Q) 0.0000900 0W QW 0.000000900 QWEI-(D
LW QW QW QW QW 0.000000900
1.000000900 000000900 QW QW 0.000000900 QW
1.000000900 0000000900 QW QW 0.000000900 QW
LWE‘KX) 0.00000900 QW 0010000900 QW QWEHD
LW 0000001900 QW QW 0010000900 0.010000900
1.000000900 000000900 QW 0.000000900 0.000000900 QW
1.000000900 QW 000000900 0000000900 000000900 000000900
1.000000900 QW 000000900 QW 001000900 0.000000900
1000000900 QW QW QW QW 0W
1.000000900 QW 0000000900 QW 000000900 QW
1000000900 QW 0.000000900 QWW 000000900 0W
1010000900 0000100900 000010900 QW 0000000900 QW
1000010900 QW 0000100900 QW 0.00000900 QW
1.000000900 QW 001000900 QW QWKX) 0.000000900
LWEM QW 0.000000900 QW 0000000900 0.000000900
1000010900 0.000000900 0.000000900 QWHX) 0.000000900 QWOE-I-(X)
1.000000900 QW 000000900 QW 0000000900 0.000000900
100000900 QW 000000900 0W 000000900 0W

 

a;

SIGNAL NAMES I

1:01:09) KOUT(10) Koo-rm) Korma)

0000100900 0W QW 0000000900 0.00000900 QW
. LW 1W 100000900 1.000000900
0.000000900 QW QWM QWOOIEKX) 0.000000900 QW-I-Q)

lWE-Ol QW‘KX) 3500000501 0.000000900 1000000901 0.000000900
QMOOQE-HX) QW 5111111190! QW 0000000900 0.000000900
0.000000900 QW 8300000901 QW 0010010900 0.000000900
. . 8W0! QW 0.000000900 0.000000900
QWEM QW 8W0! QW 0.000000900 QW
0000000900 0.000000900 8.5m-Ol 0001000900 QWEKD QW
QWEM QW 8W0! QW 0.000000900 QW
0.000000900 0.000000900 8W0! QW 0000000900 0010000900
QW QW-I-(X) 8M0! 0.000000900 QWEAX) 0000000900
0.000000900 0000000900 8W0! 0.000000900 QWEW 0.000000900
QW 0000100900 BM-OI 0.000000900 0.000000900 QW
0.000000900 QW 8W0! QW'KX) 0.000000900 QW
QW 0000000900 8M0! 0.000000900 QWE-t-Q) 0.000000900
0.000000900 0000000900 850M-Ol 0.00000900 QW 0.000000900
. 8500000501 0.000000900 0000100900 QWW
QW 0000000900 8.5m-Ol 0.000000900 0.000000900 0000100900
0.000000900 QW 8M0! 0.000000900 0000100900 QW
QW QW 8W0! 0.000000900 0010000900 QWEHX)
QW 0000000900 8W0! 000000900 0.00000900 0W
0000010900 0000000900 8W1 000100900 0.000000900 0000100900

§
i

I
I

I

Figure 29: Results of Low Gain Sigmoid Neuron Model.

76

 

 

 

 

Figure 29. (continued).

77

11MB L SIGNAL NAMES I
I
CNS.) I K001‘(13) K00'1‘(14) K001‘(15) KOUT(16) K0010?) K001' (18)

‘ ' 0 'I 0000000900 0000000900 Q W 0000000900 000000900 QW
4 ! 1.000000900 1010000900 1 .000000900 111me 100000900 1.000000900
14 I 000000900 0.000000900 Q WEI-IX) 0.00000900 QW 0.000000900
24 I 1000000900 LW 1.000000900 100000900 LW LWEHX)
34 I OWN!) QW Q 000100900 0W QW 0000010900
44 I 5WD! SW-Ol SHINE-01 8W1 8M0! SWOI
54 ! 0000000900 QW 0000000900 000000900 QW 0010010900
64 I OWEN!) 1W1 0000000900 9100000901 1W! 1501000901
74 I OWN QW 0W QWE‘IQ) 0.000000900 0.000000900
84 I Q 000000900 1W1 Q 00000900 WE—Ol “HMS-01 lWB-Ol
94 I 0. W Q 000000900 Q 000000900 0011000844!) QW 0.000000900

104 I Q WEI-(X) 15WE-01 Q W14!) 8W0] 1W! 1.5IXXXJOE-01
114 I Q 000000900 0 .000000900 000000900 QW 0.000000900 QW
124 I QW 1500000201 0000000900 8M0! 5.WOE-01 15li
134 I 0.000000900 0000000900 0010000900 QW 0.00000900 0010000900
144 I QW 1W0! QW 8W0! 5000000901 1.5(XXIOOE-01
154 I 0.000000900 Q 000000900 QW'IN QW QWKX) 0010000900
164 I QWOOOEM 1W0! QWE'I-Q) 8W0! 5.WOE~01 15000001501
174 0.000000900 Q WEI-(X) 001000900 0.000000900 0.000000900 QWEHD
184 I QW IMO! QW 8m-Ol 5.IXXXIX)E-01 1.5(XXJO0E-0l
194 I QIXXXDOE-I-(X! Q 000000900 0.00000900 QW 0.000000900 0.000000900
204 I 0.000000900 1500100501 0000000900 8M0! SWINE-01 1500000901
0W 0.000000900 000000900 QW 0W 0.000000900
11MB I SIGNAL NAMES I
(NS? I K001'(l9) KOUI‘QO) K001 (21) K001(22) K001‘(23) K001 (24)
0 I QWEM 0000000900 QW 0000000900 0010000900 0000000900
4 I LWOEHX) 1.000000900 LW 1W 1000000900 LWEHX)
14 I 0000000900 QW 0000000900 0000000900 QW QWEW
24 I 1000000900 LW LWW 1.00000900 MINNIE-14!) 1.000000900
34 I QWE-I-Q) 0.000000900 0010000900 0.000000900 QW 0.000000900
44 I LIXDOIXJEM LWEI-(D LWB‘KX) 1.000000900 LW 1000010900
54 I 0.000000900 Q W 0.000100900 Q 00000900 0000000900 QWEW
64 I LWEAXI 1 0100001900 1000000900 100000900 LWEKX) LWEHX)
74 I 0.000000900 0. 000000900 Q W44!) 000000900 0000000900 0.“!!me
84 I 1000000900 1.000000900 1000000900 1.000000900 1.000000900 1000010900
94 I OWEN!) Q WEI-(X) Q 000010900 0000000900 QW QWEI-(X)
104 I LW 1.000000900 LIXXXXXEKX) 1.000000900 HUME-(X) LIIXIOOOEAX)
114 I QWE-Hl) 0.000000900 0.000000900 0.000000900 0.000000900 01!!!)me
124 I 1001000900 1.000000900 1.000000900 LW LWXXIEM l.QXDOOEHD
134 I Q 000000900 0.000000900 0.000000900 QQXXJOOW 0.000000900 0.000000900
144 I LWENX) LW 1.00000900 1.000000900 1.0)0000900 l.QXXXXIEHX)
154 I Q “DODGE-IQ) 0.000000900 0000000900 0.000000900 0000000900 0.(XIXXX)E+(D
164 I 1010000900 1000000900 LW-I-(XI 1.000000900 1.000000900 1000010900
174 I 0.IXXX)00E+IXI 0 010000900 000000900 QWEHI) Q WEI-(X) 0000010900
184 I 1.000000900 LWEM 1.00000900 l.QXXIOOE-I-IX) LWBM LIDOOOOEHD
194 I Q MOB-14!) QWM 0000000900 0.000000900 QWEI-(X) 0.000000900
204 I 1.WE+N 1W4“) 1000001900 MIME-14!) 1.000000900 1.0)0000900
214 I Q 000000900 000000900 000010900 QW Q WEI-(X) QW

It!
a

%

11MB!—

1
(NS) I

HHHHHF‘I‘I—‘HH

fissasssrs

O

greasrsraz~

:

3

I”:
a

reassessssffifiirtrfiie°_

§HH--HHH~

K001'(25)

0000000900
1 .(XXXXXIEHD
0. WEI-(X)
1 .000000900
0. WEI-(K)
LWEM
QWEM
1.000000900
0.000000900
1.0(XXXX)E+(X)
0000000900

LIXXIOOOEW
QWOOOE-I-IX)
LOIDOOOE-I-(X)
QOIXIOOOEW
LIXXXIOOEKXI
QIXXIOOOEHX)

l.QDOIXIE-I-(X)
QWOOOEM
1.000000900
Q WEAK)

Q 000000900

K001'(3 1)

QWEHXI
LWOEHX!
QWEKX)
LWBM
QIXJOOQJEM
LIXXXXXIEAX)
0.000000900
1.000000900
QWEHX)
1.000000900
QWEM
”HINGE-14X)
QWOOOEAX)
1.000000900
QWEM
1.000000900
0010000900
1001000900
0.000000900
l.QXDOOEM
QWEM
1010000900
QWEHX)

0000000900

K001' (32)

K001 (27)

Q W
1 .W
0. MOE-14X)
1 .000000900
Q WEI-(X)
1000010900
Q W40)
1000000900
0.000000900
1010000900
0000010900

1.000000900
0.000000900

100000900
0.00000900
LW
000000900
1.000000900
QWKX)
1000000900
0.000000900
1000000900
000000900

K001 (33)

Q W
1 .W
01!!!me
1.000000900
0.000000900
1.000000900
0.000000900
LWOE‘KXI
QWE-HX)
1000010900
0000000900
1.00000900
0.000000900
1.000000900
0010000900
101000900
0.000000900
l.QXXXXE'I-(X)
0000000900
1.00000900
0.000000900
1 .W
Q 01000900

SIGNAL NAMES

SIGNAL NAMES

[(00104)
Q W

K001‘(28)

Q W
1000000900
Q 000000900
1 00000900
000000900
1000000900
QWM
1.000000900
0.000000900
1000000900
QIXIOOOIE+00
MIME-(X)
QW
LIXXXXXIE-I-(X)
0000000900
1.000000900
0.000000900
1.000000900
QOIXXIOOE+00
1.000000900
QW
LWEW
0000000900

1W
0 .000000900
1.00000900
0.000000900
1.000000900
Q 00000900
1.00000900
0 .00000900
1.00000900

Figure 29. (continued).

78

KW)

K001‘(35)

0.000000900
1001001900
0000000900
LW
0000000900
1.000000900
Q WEI-(X)
1 .000000900
QWOOOE-I-Q)
1.000000900
0000000900
1.000000900
QWE'I'Q)
1.000000900
0000000900
1.000000900
QWE-I-(X)
1.000000900
OWE-14!!
LWEAX)
0.000000900
1000000900
0001001900

1000000900
Q W44!)
111!wa
0.000000900
MINNIE-14X)
QW
1.000000900
QIXXXIOOE-I-Q)
1.000000900
QW
LIXXXXDEHX)
QWEM
1000000900
QWEHX)
1.000000900
QWEI-OO
1.000000900
0.000000900
LWE'HX!
0000000900
1000000900
0000000900

K001'(30)

0000000900
LWEM
0000010900
”KIWI-2+“)
QWEHX)
LWEHX)
0.(XJ(XXX)E+00
1.000000900
0.000000900
1.000000900
QIXXXXDE4-00
1010000900
Q WEAK)
1 .010000900
0. WEI-Q)
1.000000900
QQXXXIOEHD
1.000000900
QW
1.000000900
Q WEI-(X!
1.000000900
QW

K00'I‘(36)
0. 000100900 Q WEI-IX)

1 .W
0.000000900
1.000000900
QWEH'X)
l.QXXXDEI-IX)
QWE+N
1.000000900
QIXIOIXXJEM
LIXXXXDE'HX)
0.000000900
l.QXXXIOEHX)
0.Q)0(X)OB+IX)
”HIDDEN”
QWE-I-Q)
1.000000900
0.0!”me
14!“)me
QWEHX)
1000010900
QWEHD
1.000000900
QW

 

 

 

 

Figure 29. (continued).

79

11MB I SIGNAL NAMES I
KOUT(37) KOUT(38) K001 (39) KO0T(40) [(001141) K001 (42)
I
0 l 0000000900 0000000900 QW 0. W 000000900 QW
4 I l.QXXIOOEHX) 1000000900 LW 1000100900 1W 100000900
14 I 0.000000900 0000010900 QWEKX! 0000000900 QWM 0000010900
24 I l.QXIOIXIEI-(X) l.QXJIXXIE-I-Q) LWEM 101000900 1000000900 1.000000900
34 I QWEKX) 0.000000900 0010000900 0000100900 QW 0.000000900
44 I 1.000000900 1.000000900 1.000000900 1.000000900 1010000900 ”DOME-0Q)
54 I 0.(XXXXXIE+IXI 0.000000900 0010000900 QWHD 0010000900 0000010900
64 I 1001000900 LW 1000000900 LWW LW 1.000000900
74 I QIXIOOIXIEAX) 0.000000900 Q 44!) 0.000000900 QWOB-I-IX) QWEKX)
84 I 1.0(XIOIXIE+(X! 1.000000900 1. 900 1.000000900 1.000000900 LIXXXIDEM
94 I QWEM 0.000000900 0. B44!) 0.000000900 0.000000900 QWEHX)
104 I l.QXJOOOE4-(X) MIME-14!) 1 +00 LW 1000010900 1010000900
114 I 0.000000900 0000100900 0.000000900 QWE-I-(X) 0.000000900 QWOOOEW
124 I l.QIIOOOE-I-IXJ 1.000000900 LWKX) 1.000000900 1.000000900 LW
134 I 0.000000900 0000000900 QW QW 0.000000900 QW
144 I 1.000000900 1.000000900 1.000000900 l.QXXXJOE-I-Q) 1.000000900 1000100900
154 I QIXDOOOEHX) 0000100900 0000001900 QW 0000010900 QW
164 I 1.000000900 1.000000900 LWW 1.000000900 1010010900 1111110054!)
174 0.000000900 QWE‘KX) 0.000000900 QW QWIKXI 0.000000900
LIXXIOOOBKX) 1000000900 1000001900 l.QXXXJOE-I-(XI 1.000000900 l.QXXDOEI-Q)
QOQXJOOEI-(X) 0.000000900 Q 00000900 0.000000900 Q WEI-(XI QWOOOEHX)
l.QDOOOEI-IXI LWE-I-(X) 1000000900 1.000000900 1WE4G) 1 .QXXIOOE-HII
0001000900 0000000900 0000000900 QW 0000010900 Q 010000900
TIME I SIGNAL NAMES I
I
(NS) I KO0T(43) K001 (44) K001 (45) KOUT (46) K001 (47) K001 (48)
0 I WAX) 0.00000900 Q W 0W 000000900 QW
4 I 1.000000900 1.000000900 1 .000000900 LIXXXXDEHD 100000900 100000900
14 I QWEKX) QWE-I-(X) 0 .000000900 0000000900 QWENXI 0010010900
24 I 1.000000900 ”WINE-(X) LW‘I-(XI 1.000000900 1.000000900 LWEAX)
34 I 0.000000900 0.000000900 QWE'IN 0.000000900 0000000900 QWEM
44 I LIXXXIOOE+OO l.QWOOEd-(X) 1.000000900 1.000000900 1.000000900 1.000000900
54 I QMXIOOEI-(X) QWEMD 0.000000900 0.000000900 0.000000900 QOIXXXDE-I-(X)
64 I l.QDOWE't-Q) 1.000000900 1.000000900 LIXXJOIXEAX) 1.000000900 1.000000900
74 I 0.000000900 QWEW 0000100900 0.000000900 0000000900 QWEHD
84 I 1.000000900 1.000000900 LWE't-(X) LIXXXIOIBMX) LWE-I-(X! 1.000000900
94 I 0.000000900 QW QW-I-(X) 0000000900 QM“ 0.000000900
104 I 11111100514!) LNOQXJEM 1000000900 LW 1010000900 l.QXXXDEHD
114 I QWEM 0.000000900 000000900 Q W QWE-I-(X) 0010000900
124 I LWOEH'X) l.QXXXXIE-I-OO LWW 1 .000000900 LWEKXI 1.000000900
134 I 0.000000900 0.000000900 QWHXI Q WEI-(X) QWE‘KX) QWEAX)
144 I 1.000000900 1.000000900 LWEHD 1.000000900 1.000000900 l.QXXIOOEM
154 I 0.000000900 0010000900 0.000000900 0.000000900 QWEM QWOEAD
164 I 1.000000900 10!!meva LWM 1.000000900 l.QXXXDB-Q) l.Q!(XJOOEAXI
174 I 0.000000900 0.000000900 000000900 0.000000900 000000900 QWEHX)
184 I l.QXXXIOE-I-IX) MINNIE-14X) “KIM-100 ”INDOOR-14X) 1000000900 1 .QXXIOOB-I-IX)
194 I 0.000000900 MIME-14X) 0.00000900 QWBW Q WEI-(X! Q WEI-(D
204 I 1010000900 1000000900 1.000000900 l.Q!(DOOE-I-IX! 1000010900 1.000000900
214 I QWEM 0000100900 QWE‘I-(X) QW Q 00000900 QW

r§““

ES”

 

113.15 I—— —SIGNAL NAMES '
(NSI) I KOUTO) KOUTQ) KOUTG) KOUT(4) KOU'KS) KOUT(6)

I
E
E
s
I
I

E§§§§§EEESEEEEE‘

I
I
I
I

I

I

I

I

_§_g
I
I

KOUTO 1) KOUTOZ)

22:-extreme

I
I
5
I
I

Figure 30: Results of High Gain Sigmoid Neuron Model.

80

aoooooomoo o.oooooo£+oo 0.000000%» onoooooawo omoooaawo 0.0000005
l(iIooooooIatgci) Iiaiooooafl) l(iooooooewo 10' +00 1° +00 1' :83
o. E+00 0: +00 .oooooomoo o.ooooooa+oo o.oooooos+oo

TIMBI

gffliirtrfixtg

Ktfifii

fifiifitfifi

§§§§E§

SIGNAL NAMES I
I
(NS) I KOUTOB) K0011“) KOUT (15) KOUTOG) KOUTO7) KOUTOS)
I
QWEHXI o.oooooo£+oo o.ooooooa-oo 090000015400 DWI-(X) QWE-o-m
1300000300 1.0000005+00 Loooooos+oo 1.ooooooa+oo Looooooa+oo l.QXJOIXIE-I-(X)
QWEd-(XI o.oooooor-:+oo QWEM o.oooooas+oo o.oooooos+oo QWEKXI
1.moooo£+oo LIXXXXJOBm 1.ooooooE+oo Looooooe+oo l.QXXIOOEm 1.ooooooE+oo
o.ooooooa+oo o.oooooos+oo o.ooooooa+oo Qmm 00000009300 o.ooooool-:+oo
QWERXI o.ooooooE+oo 09000001900 woooooaoo QWEHX) QWEHX)
QWEKD QWEI-(l) QWE'I-(X) LWIEM o.oooooos+oo QWEW
o.ooooool=.+oo o.oooooos+oo QWAXI Loooooomoo QWOOOE-I-(II QWEI-(X)
o.oooooon+oo QWE-o-m 0000000300 LWM o.ooooooa+oo QNOOQIE-o-(X)
QWE'I-w o.ooooooE+oo QWM LW-o-m QW 00000001900
ooooooomoo o.owoooI-:+oo o.oooooos+oo LW-I-(D QQXXJOOE'HXI o.oooooos+oo
o.oowoo£+oo o.ooooooe+oo o.oooooa-:+oo LW o.ooowon+oo 03000005300
QWEHX) 0900000900 o.oooooos+oo LWW QWE+OO QQXXJOOEAX)
o.oooooos+oo o.ooooooa+oo o.oooooa=.+oo LWEHX) 0.0000(an o.oooooos+oo
QMXIOOEM o.ooooooa+oo o.ooooom=.+oo LW o.ooooooe+oo o.ooooooE+oo
o.oooooor-:+oo omoooomoo oooooooaoo LWOEM 00000005400 03000005400
QW 0mm o.oooooua+oo LW 0900000300 0.00000054-00
I SIGNAL NAMES I
KOUT(19) KOUTm) KOU'I' (21) KOUTC22) KOU'I'(23) KOUTQA)
QWEM o.oooooos+oo QWE-o-m 0W amour-2m o.ooooooa+oo
10000001900 10000001900 Loooooomoo 10000001900 11100000300 LIXDOIXIEAX)
QWEI-Q) o.ooooooE+oo o.oooooos+oo onooooo£+oo QW QWEM
1mm Loooooos+oo Locomomoo l.QXIOIXE-I-m LW Loooooomoo
QM“ 0.000000134-00 09000001900 o.oooooo£+oo QW o.ooooooE+oo
IWEM l.QXXXIOE-o-(D Loooooomoo Loooooos+oo MINDS-RX) Loooooomoo
QWEM Qmm QWE‘KX) o.oooooo£+oo QQXXJOOEI-(X) o.ooooooa+oo
o.oooooos+oo QWEW QMXXJOE-O-(XI omoooomoo LW 0.00wooE+oo
QWM QW 09000003410 00000005400 LW o.ooooooa+oo
QWEd-(XI o.oooooors+oo QWEd-(l) o.ooooooE+oo 10000005400 QWEM
QWE-I-(X) o.oooooos+oo QWBW QW'I-(XI 1.ooooooE+oo 0.0000005+oo
o.oooooos+oo o.ooooooa+oo o.oooooaa+oo o.oooooos+oo Loooooo£+oo omoooomoo
o.ooooooa+oo QWEM QWM QW LWBM o.oooooos+oo
o.oooooor-:+oo o.oooooos+oo o.ooooooe+oo 0.00M” Loooooomoo 0.(XX)000E+Q)
o.omooor.+oo 09000005400 ammo o.oooooos+oo Looooooa+oo QWEHX)
QWEM o.ooooom=.+oo 0900000900 (100000013400 10000001900 00000001900
00000005300 ooooooomoo omoooa-zwo QW 1900mm QW

 

 

 

 

Figure 30. (continued).

81

TIME I

 

 

 

 

Figure 30. (continued).

82

SIGNAL NAMBE I
I
(NS) I KOUTQS) KOUI‘OG) KOUT(27) KOUTQS) KOUT(29) KOUTGO)
0 I amouaoo 0.0mm Q W 0W 0W aoooooomoo
4 I Lamoswo LW 1 .W Looooooe+oo Locoootmoo LWEHXI
14 I omnooomoo 0000000900 Q ooooooa+oo QOQIOOOEM QWOOOW QOOOOIDEW
24 I 10000001900 Locomwo 19000001900 LWAX) l.QXXXXJEm LWEM
34 I o.ooooooa+oo QW 09000001900 QWd-(XI QW 0000000900
44 I LWE'IN Looooooa+oo LW 10000001900 LW LWEKX)
54 I QWEM o.omoooe-oo oooooooewo o.oooooos+oo QW o.ooooool=.+oo
64 I MIME-04X) LW 1.000000£+oo l.QXXXXE-I-(X) kW LWM
74 I o.oooooos+oo QIXIXIOOEAXI o.oooooos+oo o.oooooos+oo 0.0(XXXXIE+(X) o.oooomE+oo
84 I o.oooooo£+oo LW oooooooaoo o.oooooaa+oo QWEi-N 09000005400
94 I o.oooooo£+oo Loooooomoo o.ooooooa+oo o.ooooooe-oo 0900000300 09000001900
104 I o.oooooox=.+oo 1.000000%» o.oooooa=.+oo QWEM o.oooooos+oo o.oooooon~:+oo
114 I Q MOB-(X) Looooooaoo 09000003400 o.oooooos+oo o.ooooooa+oo o.ooooooa+oo
124 I Q WEI-(X) 1.oooooos+oo ammo QW 09000001900 o.ooooooE+oo
134 I Q oooooom-oo LWOEW o.oooooos+oo o.ooooooa+oo o.ooooom-:+oo o.oooooos+oo
144 I Q WEI-(X) Loooooomoo o.oooooma+oo o.oooooo£+oo Q ooooooa+oo QWE‘KXI
154 I QW Looooooewo omoooumo comma-mo Q ooooooem QW
TIME I SIGNAL NAMES I
I
KOUT (31) KOUTGZ) KOUT (33) KOU'I' (34) KOUTGS) KOUT (36)
I
0 I QWE-rm omooooaoo Q W 0W 0mm uooooooa+00
4 I kWh-(XI MINDS-om LWEM woooooewo moooooaoo Looooooe+oo
14 I ooooooomoo QWEHXI Q ooooooE+oo o.ooooooa+oo QWEHX) QWEHXI
24 I l.QXXXXIE-tm 1000000900 10000001900 hmm Looooooem 10000001900
34 I QWEHXI QW QWEAX) 0900000900 0000000900 00000001900
44 I l.lemE-KX) LW 1.000(XJOE-I-(X) 1.ooooooa+oo LW LWEM
54 I QWEAXI QW o.ooomoa+oo o.oooooos+oo o.omooor-:+oo QWE‘I-(XI
64 I Loooooos+oo LW 1 .ooooooswo Loooooamoo Looooooe+oo LWE-rm
74 I 0.000000124-00 QWEFQ) Q WEI-(X) 00000001900 o.oooooos+oo QWEM
84 I LWE‘KXI Loooooon+oo 1WE+QJ LWMXI 1.oooooor-:+oo Looooooswo
94 I QWEvI-(XI 00000005400 Q 0000008400 0000000900 o.owoooa+oo QWE‘HXI
104 I o.ooooooa+oo omoooomoo o.oooooaa+oo 0000000300 QWEHX) LWEHD
114 I QWEHX) 00000001900 o.oooooaa+oo o.oooooos+oo 0 .ooooooewo 10000001900
124 I QW o.oooooma+oo 0.000an400 0.00000054-00 0 “WEI-(X) LWOEM
134 I o.ooooooa+oo o.ooooooa+oo omoooomoo QQXDOOM Q WEI-(X) 1 .ooooooawo
144 I o.ooooooE+oo omooooswo o.oooooo£+oo QWEHX) OWE-(XI 1.00000054-00
154 l QW oooooooewo o.oooooos+oo ooooooom-oo 0mm 1 .oooooomoo

TIMBI

 

 

SIGNAL NAMES I
I
(NS) I KOUT (37) KOUTGS) Km 3001140) K0011“) KOU'I'(42)
I
0 I Q WEI-(l) omoooomoo Q W QW omoooomoo ooooooom-oo
4 I 1 .ooooooa-oo MIME-om 1 .ooooooawo LWEHX) woooooewo Locooooem
14 I OWEN QW Q ooooooewo 00000005400 QW o.ooooooa+oo
24 I 13000001900 momma-mo 1mm Loooooomoo Loooooomoo moooooewo
34 I o.oooooo£+oo o.oooooo£+oo Q W44!) o.ooooooE+oo o.ooooooa+oo QWE'IN
44 I moooooewo LWXIBM Loooooomoo LW-Im l.QXXXXIE-I-(X) LWEM
54 I OWEN!) o.oooooo£+oo Q momoo 09000001900 QW o.ooooooa+oo
64 I LWEM LW Loooooosm 1.ooooo(£+oo 10000001900 LWXXIE-I-m
74 I QWE'IM QW o.ooooooa+oo omooamoo 0000000300 o.oooooos+oo
84 I 1MB“) 1.ooooooa+oo MINNIE-m 1900000900 Loooooomoo LWEAXI
94 I OWE-rm QW 09000001900 omoooaswo QW QWEd-(X)
104 I LW Looooooam 1.ooooo(£+oo 1.oooooor-:+oo l.QXXXDEHD Looooooa+oo
114 I Q 0000005400 04000000900 o.oooooc£+oo (10000001900 0.0000(an QWEHD
124 I Q W o.ooooooe+oo o.oooooos+oo Loooooosa-oo o.oooooos+oo o.oooooos+oo
134 I Q MOE-HI) ooooooomoo uncommon LW 09000002400 Qm-o-(II
144 I Q WEI-(XI omoooomoo 0000mm LW omoooomoo QWEHD
154 I omoooomoo o.ooooooa+oo 0W LW omoomewo QW
TIME I SIGNAL NAMES I
I
(NS) l KOU'I‘(43) KOU'I‘(44) KOU'I' (45) ICOUT(46) KOUT(47) KOU'I‘(48)
I
0 I 0.00000054-00 0mm Q W 0W omooooa+oo QW
4 I Locomoewo Loooooomoo 1 .W mooooomoo moooooswo LWEHX)
14 I 09000001300 Q W Q WEI-(XI o.oooooa-:+oo Q W 0000000300
24 I moooooewo 1300000900 Looooooa+oo Looooooaoo mooooomoo 1.ooooooa+oo
34 I o.ooooooa+oo Q mm 0.000me o.ooooooa+oo o.oooooos+oo o.oooooor='+oo
44 I mooooomoo 1.000000%» Loooooos+oo LWE‘I-m Loooooomoo 10000001900
54 I o.ooooool-:+oo QW o.ooomoa+oo o.ooooo(£+oo QW o.oooooox-:+oo
64 I 19000001900 l.de-m MIME-(X) l.QXIOOOE-om 1.ooooom~:+oo Loooooomoo
74 I o.oooooor~:+oo 09000005410 Q ooooooe+oo o.ooooom=.+oo QW QWEKXI
84 I 10000001900 LWW LWIEM LWM Loooooos+oo Loooooomoo
94 I QWEd-(X) o.owooos+oo Q W844!) QWW 0.0000(an QWE‘KX)
104 I 1.ooooooa+oo Lomooomoo LWM 1“!me LWE-I-(X) Loooooos+oo
114 I o.oooooos+oo 09000001900 09000001900 QW o.oooooos+oo 0.00000054-00
124 I Loooooos+oo 1.ooooooa+oo Loooooos+oo hm 1.oooooos+oo l.QXJOOOEflII
134 I o.ooooooa+oo o.oooooos+oo o.oooooos+oo QW QWEM QQXXJOOEd-(XI
144 I l.QXXIOOE-rm o.ooooooe+oo o.ooooooa+oo QWEHX) o.ooooooa+oo QWEM
154 I 10000001900 000mm o.ooooom+oo QW o.oooooo£+oo o.ooooooa+oo

 

 

Figure 30. (continued).

83

TIMEI

I
(NS) I KOUTO)
I

I
é

Imwooswo

1.000000154-00
LWE-I-(X)
LW
MIME-km
LWEM
MIME-rm
”HIDDEN”
1000000900

MIME-km

Loooooomoo
IMHO-(XI
1.WE+N
19000002400
Loooooomoo

zoom)

QW
LW
ammo
MIMI-(X)
uncommon
omoootmoo
o.ooooo(£+oo
Qm-rm

KOUTG)

Q W

10000001900

0 .ooooomawo
o.oowooe+oo
QW
QW
ammo
QW
o.ooooooe+oo
QW
0W

0.0000084»
09000005400
omoooamo
o.oooooaa+oo
09000003400

SIGNAL NAMES

xoum)

o.ooooooa+oo

Looooooewo

QW
QW
o.ooooooa+oo
QW
o.ooooomE+oo
uncommon
o.ooooooE+oo
QW

QWM

o.ooooooa+oo
QW
o.ooooooa+oo
QW
QW

x001!»

QW

MIME-H!)

o.oooooos+oo
QWEI-w
ooooooomoo
o.oooooos+oo
0W
omoooomoo
00000002400
0900000900
o.ooooooe+oo

QW
0mm
omoooumo
onooooaawo
0mm

KOUT (6)

QW
1 .ooooooa-m
Q 000000500
Q WEI-(X)
Q 0000005400
QW
o.oooooos+oo

EEE’E“

§E§§E§::::::22:.

1W omoootmoo QW 0W

SIGNAL NAMEE I
KOU'l'(8) KOUTG) KOU'I‘OO)

Q W
1 .W
Q oooooomoo
LW‘IN
LWM
Looooooa+oo
LWEM
Loooooos+oo
LWBM
1 0000005400
1 9000003400
Loooooomoo
momma
Looooooa+oo
1.oooooa=.+oo
1W4“)
11111113441)

Is
I

KCXJTO 1)

0mm
tmfll)
o.oooooos+oo
o.ooooooa+oo
omoooomoo
QWHX)
ommooswo
omoooomoo
o.oooooos+oo
o.oooooa.=.+oo
ammo
o.ooooooa+oo
QW
QW
o.ooooooa+oo
aooooooawo
uoowooamo

KOUI'OZ)

QW
LW
0900000900
09000001300
omoooomoo
Q ooooooa+oo
omooooewo
0 “WEI-(X)
00000001900
omoooomoo

uoooooomoo

‘O

I

0 9000001900
Q oooooo£+oo
QWEKX)
o.ooooooI-:+oo
QWEHX)
QWEM
o.ooooooa+oo
QWE-I-(X)
QWE-I-(X)
o.oooooos+oo
QWEM
QWEHX)
o.omoool.=.+oo
QWEd-w
QW

zefifiifitfifi

Figure 31: Results of Binary Response Neuron With Capacitance.

84

MI

 

 

 

 

Figure 31. (continued).

85

SIGNAL NAMES I
I
CNS) I KOUI‘(13) KOU'I‘O4) KOU'I'OS) KOUTO6) KOU'I'(17) KOUI'(18)
I
0 I QWEM 09000001900 Q W onoooooa+oo o.oooooola+oo o.ooooooa+oo
4 I monooomoo 14000000900 1 .oooooom-oo 1 omoooaoo mooooos+oo Loooooos+oo
14 I QWEM QW Q oooooos+oo Q ooooooswo ounooomoo 0.0000005+oo
24 I 1.oooooos+oo LWW Looooooa+oo 1.oooooa-:+oo 1.0moooa+oo 10000001900
34 I 0.0000005+00 QW o.oooooor-:+oo Q oooooaa+oo Q W 00000005400
44 I QWEi-(XI o.ooooooa+oo mammal-300 Q oooooaa+oo Q WEN!) 00000001900
54 I OWE-Kl) QW 0000000300 1.oooooa=.+oo Q ooooooewo o.ooooooE+oo
64 I o.oooooos+oo o.ooooooz-:+oo ammo LW QW o.ooooooE+oo
74 I o.moooo£+oo o.ooooooa+oo ammo mooooaawo 0.0000005+oo o.oooooor-:+oo
84 I omooooaoo o.ooooomz+oo omowoswo LWI-m QW o.oooooos+oo
94 I o.ooooooE+oo o.oooooos+oo o.ooooooa+oo LWM QW o.ooooooE+oo
104 I o.oooooon+oo o.oooooo£+oo o.oooooaa+oo LW 0900000900 00000003400
114 I o.ooooooa+oo omooooawo QWM 1.00000024-00 0000000900 0000000300
124 I o.ooooooa+oo QWEI-Q) momma-“.400 l.QXXXXIEI-(X) Qm-I-m QWEKX)
134 I 00000003400 o.oooooos+oo cocoons-poo LW o.oooooos+oo QW
144 I QWOOOEHXI 09000001900 09000005410 momma-00 omoooo£+oo o.oooooos+oo
154 I QW 0W oooooouawo LW 0mm QW
TIME I SIGNAL NAMES I
I
(NS') I KOUT(19) KOU'I'QO) ROW (21) KOUT (‘22) KOU' 1123) KOUT (24)
0 I monomer-mo omooooa+oo QW QW 0mm QW
4 I 1.00000054-00 moooooa+oo LW moooooa+oo mooootmoo Loooooos+oo
14 I 09000001900 09000002400 o.ooooooa+oo o.ooooooa+oo QWEM 09000001900
24 I 1.oooooos+oo LW 1900000900 l.QXXXXE-I-(X) LW Looooooewo
34 I o.ooooooa+oo o.ooooooa+oo o.ooooooa+oo QW'IN ooooooomoo o.ooooooe+oo
44 I LWE+OO LWEAX) Loooooomoo moooooewo LW Loooooomoo
54 I o.oooooos+oo 0000000900 QWENX) QOIXIOOIEM QW QWE‘I-(X)
64 I o.oooooor-:+oo 00000001900 omoooomoo omoooomoo LW QWd-(XI
74 I 09000001900 o.ooooooa+oo onoooooaoo o.oooooos+oo LWHX) 0900000900
84 I 0900000900 QWE-I-m QWEHI) 00000001900 1111me 00000005430
94 I o,oooooos+oo QW‘o-m ooooooomoo o.ooooo(£+oo Locooooswo o.ooooooa+oo
104 I QWEAXI QWEI-N o.ooooooa+oo QW 19000008400 omooooam
114 I QWEHX) o.oooooos+oo o.ooooooa+oo QW 1.ooooooe+oo QWEM
124 I QWOOOE-I-(XI o.ooooooa+oo o.ooooooa+oo QW Looooooa+oo o.ooooooE+oo
134 I o.oooooora+oo o.oooooos+oo ammo o.oooooox~:+oo LWE‘I-(X) QW
144 I 090000013400 o.ooooooa+oo o.ooooooE+oo 0.000000%» 10000001900 o.oooooos+oo
154 I o.ooooooa+oo 0W 09000005400 QW mooooomoo QW

TIMEI

 

 

 

 

SIGNAL NAMEE I
I
(NS') I KOUT(25) KOUTGG) KOUT(27) KOU'I'OB) KOIJ'I’(29) KOU'I' (30)
0 I oooooooa+oo 0000mm Q WEI-(X) o.oooooomoo omooooewo QW
4 I 19000001300 mooooomoo 1.a)ooooe+oo moooooewo wooooomoo LW
14 I QWE-I-(XI o.oooooos+oo Q ooooooE+oo 0.0000008+oo QW o.oooom£+oo
24 I 1.(X)00(X)E+w LWEM 1.oooooos+oo Looooooawo LW Loooooomoo
34 I QWEi-(X) QW Q WEN!) Q ooooooE+oo o.oooooon+oo o.ooooooE+oo
44 I 1.oooooor-:+oo 1 .WEW 1.ooomoa+oo LIXXXXXE'I-m 10000001900 1 .WEM
54 I o.oooooor-:+oo Q W Q 000000900 Q ooooooe+oo QW Q ooooooa+oo
64 I LWEKXI 1 .ooooooewo 11!me mooooomoo MIME-rm Loooooomoo
74 I QWEKX) Q 0000003400 o.ooooooa+oo o.oooooos+oo QW Q 0000001900
84 I o.ooooooe+oo QWEM o.oooooos+oo o.oooooos+oo 00000001900 0 .WEKII
94 I omooooswo Looooooawo oooooooswo 0000000900 Q WEI-(XI 0 .WEAX)
104 I o.ooooooza+oo 1000000300 omooooewo QWE'I-(XI Q WEI-(X) 0000000900
114 I QWOOOEND LWE-I-m QWEI-M o.ooooooE+oo Q ooooooewo QWEHX)
124 I QWEAX) 1.ooooooa+oo o.oooooaa+oo QWEHXI o.ooooooz-:+oo o.ooooooa+oo
134 I QWOOOEHX) LW o.ooooooa+oo QWE-o-(XI o.ooooooa+oo omoooomoo
144 I o.oooooo:—:+oo mooooomoo o.oooooaa+oo MIME-om oooooooawo QW
154 I QWEHX) 1mm 0mm QW QWd-(D 0.00000054-00
TIME I SIGNAL NAMES I
I
(NS) I KOUT(31) KOUI‘(32) XOUT (33) W) KOUI‘OS) KOUTG6)
I
0 I o.ooooooe+oo QM“ QW 090mm onoooormoo QW
4 I Loooooom-oo LWEM LWE+QI 1W woman-mo LW
14 I OMB-(X) QWEHXI 09000005430 o.oooooo£+oo QW 09000001900
24 I kWh-(XI LW LWEM Looooooawo LW Loooooomoo
34 I OWEN!) Qme 00000001900 0000000900 QW 0000000900
44 I LOWER!) 11!!!)me Looooooa+oo 1000000500 LW l.QXXXXIE-o-m
54 I 00000001900 QW 00000005400 omoooamo QW QWEM
64 I IWEKI) Looooooa+oo Looooooa+oo 1.ooooou-:+oo Locooooa+oo MIME-om
74 I o.oooooos+oo QWEAX) Q WEI-(X) o.oooooos+oo QW 0000000900
84 I 1.oooooor-:+oo 1.00000054-00 MIME-om Loooooos+oo LW Loooooos+oo
94 I QWEI-Q) Q 000000300 o.ooooooa+oo 0900000900 QW ammo
104 I QWOOOE-I-(XI Q WEI-(XI o.ooooooe-oo ooooooo5+oo 00000001900 1.0000005m
114 I QWOOOEd-(XI Q oooooomoo o.oooooos+oo QW QWE-I-(XI Loooooomoo
124 I QMOOOEHX) 0 .oooooomoo o.oooooaa+oo QW onooooomoo MINNIE-m
134 I QWOOOFA-(XI o.oooooos+oo o.oooooa=.+oo QWE‘I-(XI 0.IXXXIQIE+(X) 1.oooooos+oo
144 I QWE-KX) o.ooooooe+oo o.ooooooa+oo QWEi-(XI Q WEI-(XI 1300000300
154 I QWEM o.ooooooa+oo 0W QW ammo hm“)

Figure 31. (continued).

86

iiifiififsssssssszs-E—I

I
r

 

 

 

=:s_§_§

fil‘iitffi

§E§§E§

 

Figure 31. (continued).

87

SIGNAL NAMES I
KOUT(37) KOUI'(38) KOUT (39) KOUT(40) [001141) KOUT(42)
QWE‘I-(XI ooooooomoo QW 0W o.ooooo<£+oo QW
LQXXIOOE‘o-m Loooooom-oo LW 1W mooooomoo LW
OWE-Q) 0.00000054-00 o.oooooos+oo o.oooooa-:+oo 0.(XXX)OOE+OO 000000015400
l.QXIQXJE-I-(X) Looooooa+oo Loooooos+oo Looooooa+oo 1.ooooool-:+oo LWE-I-m
QWEKXI QW o.wwooE+oo 0900001300 QW o.oooooos+oo
LIDOOIXIEM LW moooooawo 1 .ooooormoo Loooooomoo l.QXXXXIEI-(XI
00000001900 QW o.oooooon+oo Q oooooos+oo QW QWE‘KXI
LIXXIOCDE-I-m l.Q!XXXIE-I-m 19000001900 1 .oooooaa-mo Looooooawo MIME-(X)
WEI-(X) QW oooooooawo Q oooooomoo o.oooooo:-:+oo QWE'IN
Loooooomoo l.QXXXIOE-KX) LIXXXXXIEM l.Q!)OOE-om Looooooa+oo hm!”
o.ooooooa+oo QW o.ooomos+oo o.oooooaa+oo o.oooooos+oo QWEKXI
1 .MOEM 1000000500 1W Loooooow mooooomoo 1.0000(an
Q ooooooa+oo omooooswo o.oooooaa+oo QW 09000002400 QW
Q oooooo£+oo o.ooooooa+oo ammo LOWE-rm o.ooooooe+oo 0300000900
Q WEI-(X) o.ooooooa+oo o.oooooa-:+oo LW QWEd-m QWEW
QWE-I-m omooooaoo o.oooooua+oo LWEM Q oooooomoo 03000001900
OW 0mm OW LW oooooooawo QM“)
I SIGNAL NAMES I
XOU'I'(43) K0011“) KOUT(4S) KOUT (46) KOUT(47) KOUT (48)
o.oooooor~:+oo o.ooooooa+oo Q W OW omooocmoo QW
1.ooooooI-:+oo mooooomoo 1W 1W moooooewo l.QXXXIOEm
o.oooooos+oo QW omoooomoo o.ooooooa+oo QWEM QWEKX)
mooooomoo LW 1W Looooooewo 10000001900 10000005400
QWEi-(X) QW onoooooewo 0900000900 o.ooooooa+oo o.ooooooa+oo
LWEM 1 .ooooooam IWEW moooooewo LW Loooooomoo
o.ooooooE+oo Q WEI-(XI 09000005400 o.ooooom-:+oo QWEM QWEI-Q)
1.0(XIOQIE-I-(X) 11!me Loooooomoo l.QXIOOIE-I-(X) l.QXIOMEI-m 1.oooooos+oo
QWEd-(X) o.ooooooa+oo ooooooos+oo o.oooooo!-:+oo o.oooooos+oo omooooewo
Looooooewo LWEAX) mooooomoo 1.oooooaa+oo 1.00000054-00 10000001900
QWEHXJ o.omoooa+oo oooooooam o.ooooom-:+oo QW onoooooa+oo
l.QXJOOOE-I-(X) moooooawo 1.ooooom-:+oo 1.oooooo£+oo l.QXXImE-o-(I) l.Q!)(XIOEm
QWEHXI onoooooa+oo o.oooooos+oo QW QQX'IOWEHX) 03000005400
1.00000054-00 looooooawo 10000005400 Looooooa+oo l.QXXXXIEd-m l.QXXXIOE-I-m
QQXXIOOEi-W 00000005400 ooooooos+oo 0900000900 QWEM o.oooooos+oo
1.00!an 0300000900 ooooooomoo QWE-I-m QWEW 03000001900
1W omoooomoo ammo QW QWEHXI 00000005400

 

113:3 I SIGNAL NAMES I
(NS') I KOUI‘O) KOUT(2) KOUTG) KOUT(4) KOU'I'(5) KOU'I‘(6)

OI onoooooswo OW QW o.oooooos+oo QW o.oooooos+oo
I ISM-01 ISM-01 ISM-OI 7M~Ol 75WE01 7WE-01
10000001900 oooooooswo 0 .W O. W Q W4!) Q W
1 .WXXIE-I-(X) 0mm omoooomoo Q WEI-(X) omoooos+oo Q W
LWEHX) 0.000me QW o.ooooooa+oo o.oooooos+oo QWEHX)
l.QXXXXIE-I-m o.oooooaa+oo 0W 0000000900 QWEIN QW
LW omoooosm o.oooooos+oo Q WEI-(X) o.oooooos+oo QW
LW o.oooooa-:+oo QW Q W QWEM 00000009400
LWEW ammonia-+00 ooooooos+oo Q W QWM 030000013400
Looooooa+oo 0W OW Q WEI-(X) 00000005400 QW
LW omooouawo 0000000500 QW o.oooooos+oo QW
looooooawo QW o.ooooo(£+oo o.oooooo£+oo o.oooooaa+oo QW
1.ooooool~:+oo o.oooooos+oo o.ooooool:-.+oo o.oooooos+oo 0900000900 o.oooooos+oo
1.0000005+oo QW omoooaawo QWE'HX) Q oooooos+oo o.oooooor-:+oo
MIME-04X) QW o.oooooo£+oo QWE-I-(X) Q ooooocmoo 0.(XXXXJOE+(X)
1.oooooo£+oo QW QMW 09000001900 0 .ooooowwo OWE-(X)
1.oooooos+oo QW o.oooooa~:+oo o.ooooooa+oo O .oooooomoo QW
10000001900 QW o.oooooos+oo QWEHX) 090000015400 o.oooooos+oo
19000001900 QW o.oooooos+oo o.oooooos+oo o.ooooooI-:+oo QQXIIXIOE'I-(X)
Loooooos+oo QW omoooamo o.ooooooE+oo 09000005400 03000005400
10000001900 QW o.oooooa-:+oo o.ooooooE+oo o.oooooa-:+oo 0.00000054-00
1W4!) QW Q oooooamo 0900000900 o.oooooox-:+oo o.oooooos+oo
lsoooooswo QW ammo QW o.oooooa.=.+oo OW

§§§§§§E§§§§§g§ggggg§.

:3
A»

SIGNAL NAMES I
K001“) K0070) KOUT (10) KOUTO I) KOUI‘(12)

o.ooooool=.+oo 0mm 0. W Q W 0900mm QW
1500000501 7W1 ISM-01 ISM-01 7W1 7500000501
0000000900 QW OW Q ooooooa+oo QW O .oooomewo
m-Ol OW swam-01 Q oooooos+oo ISM-01 Q oooooos+oo
OWEN!) QW momma-01 Q WEI-(XI o.oooooos+oo Q W
09000001900 QW SWE-OI QWOOOE-I-IXI QWE-IQ) o.oooooos+oo
omoooomoo QW 7500000501 omoooos+oo 0900000500 00000005400
o.oooooo£+oo o.oooooos+oo LIME-01 QW 0.000000st o.oooooos+oo
00000005400 QW 7500000201 ooooooom-oo ooooooos+oo QWE-o-Q)
00000001900 QW ISM-01 o.oooooom o.oooooos+oo o.ooooooa+oo
QWE-I-(X) o.oooooos+oo ISM-01 QWEd-N o.oooooor-:+oo QWE-I-(X)

I-I
I

ssssgtr§:.o_

104 I QWEHD o.oooooos+oo ISM-01 QWE-o-Q) o.oooooos+oo QW
114 I QWEI-Q) o.oooooos+oo ISM-01 o.oooooos+oo o.oooooos+oo QOIXXXIOE-I-(XI
124 I QWEHX) ooooooomoo ISM-OI o.oooooos+oo QWE-I-(X) QW
134 I QW OWN!) 7500mm o.oooooos+oo QW QW
144 I o.oooooos+oo omoomswo ISM-01 o.oooooos+oo QW omooooa-oo
154 I o.oooooos+oo o.oooooos+oo vsooooaa-m o.oooooos+oo 0.000000154-00 o.oooooos+oo
164 I o.ooooooa+oo o.oooooos+oo ISM-01 QWEHX) QW QW
174 I o.oooooos+oo OWE-«(XI ISM-01 QWE-KX) 0.000000%!) 0.00mooE+oo
184 I Q W 000000012400 LIME-01 omoooos+oo QW QW
194 I Q WEI-(XI 0mm ISM-01 09000002400 doom-+00 o.oooooos+oo
a): : QW QWE‘KX) ISM-01 Q WEI-(XI o.oooooos+oo o.oooooos+oo

QW 0W ISM-01 omooooswo o.oooooos+oo onooooomoo

Figure 32: Results of Low Gain Ramp Response Neuron With Capacitance.

88

§§§§§§§§§E§93393t2§=:3_3

I‘.’
;

;.° 3

§§§§§§§E§§E§ffifltr§

v-

[00103)

10.11114)
omoooosm

7W1

QW
SW1
o.oooooos+oo
MOI
QW
2W1
Q W
Q oooooom-oo
Q oooooos+oo
Q WEI-(X)
0900000300
0900000300
09000001900
o.oooooos+oo
o.oooooos+oo
QWEAX)
090000013400
osooooos+oo
0.000000st
omoooos+oo
0mm

omoooomoo

' SHINE-01

SIGNAL NAMBE J.
KOUTOS) KOUTO6) K0011”) KOU1'(18)

Q W OW 0W Q W
ISM-01 ISM-01 1500000501

0000000900 uncommon QW ooooooos+oo
SWINE-01 1W! smoooos-m soooooos-on

uncommon o.oooooaa+oo ooooooos+oo o.ooooooE+oo
ISM-OI SWINE-01 SW-Ol ”WE-01

o.ooooooa+oo o.oooooaa+oo o.oooooor-:+oo 0000000300
0000000900 1000000501 ISM-01 WE-Ol
o.oooooos+oo ISM-01 o.oooooos+oo o.oooooos+oo
o.oooooos+oo 1mm W401 QWE-I-Q)
09000002400 “HEM-01 QWEHD
ammo 1W]
o.oooooas+oo ISM-OI
QWM ISM-01
o.ooooooe+oo ISM-01
090000013400 ISM-OI
QW 7W1
090000th 7W1
omooocmoo 7W1
o.oooooaa+oo 7W1
omooooseoo ISM-01
ammo 7W1
OW 7W1

SIGNAL NAMES I

KOUI'(21) Km) X00103) KOIJ1‘(24)

Q W OW 0mm QW
ISM-01 ISM-01 ISM-01 7W1
O. oooooos+oo QWEW QW 0000000300
5. MIKE-01 5. WIDE-01 EMS-01 5.(XXXXXIE-01
Q WEI-(XI Q oooooas+oo O. oooooos+oo Q WEI-(XI
1000000901 10000001501 5 .WE-Ol 5 .(XXIDOE-Ol
omooooswo o.oooooo£+oo QW o.oooooo£+oo
“HIRE-01 5.QX)OOOE-OI QWE-Ol 5.(XXXXXIE-Ol
o.oooooos+oo o.ooooooa+oo QW 090000015400
1W0] QWEOI 5.(XXXXIIE~01 QWE-Ol
0.000000st omoooomoo o.oooooos+oo o.oooooos+oo
5 .(XXXXIOE-Ol QWE-Ol someone-01
Q oooooaa+oo Q W omooooswo ammonia-+00
ISM-01 zsooooos-ox someone-01 W01

Q oooooos+oo
ISM-OI
o.oooooos+oo
ISM-01
o.oooooa=.+oo
ISM-01
o.oooooos+oo
ISM-01
OW

o.oooooos+oo
M01
QWE‘I-Q)
2500000301
o.oooooos+oo
LIME-01
ooooooos+oo
LIME-01
ooooooos+oo

Figure 32. (continued).

89

LIME-01

SWE-Ol

WE-Ol

SOME-01

LIME-01

LWE-Ol

ISM-01

SWE—Ol

1500000501

Q oooooos+oo
lSWE—Ol
Q WEI-(XI
WE-Ol
o.oooooos+oo
2.5(IXXJOE-Ol
o.ooooooa+oo
WE-Ol
QW

I

2§§iiiiiisafirrrer:2:::g-§

3-;

I:
&

§§§§§§§§§E§39§35¢trfizeo

1P

KOUTQS) m [001(27)

0000000900

KOU1(31)

000000m-:+00
7W0!
000000013+00
SHINE-01
00000005400
5000000501
Q WEI-(X)

SIGNAL NAMEE I

 

xoumx) tom-(29) KOUTGO)

OW QW OW OW QW

7W1

7W1 IMO]

SIGNAL mums g
m tomes) xom‘cu)

 

£00165) was)

0W QM!” OW OW QW

7500000501

7m-Ol

7300000201 7500000501 7500000501
000000aa+00 0mm 0.00moos+00

5000000501 5000000501 50000001301
0000001900 00000005400 0000000300

5000000501 5000000501 5000000501

QW QW QW
5000000501 5000000501 5000000501

Figure 32. (continued).

11MB I SIGNAL NAMES I
I
018') I K00 1137) K0011”) K0011”) K001140) K0011“) X00132)

QW omoooos+oo QW 0W 0W QW
7500000501 7500000301 1.5000050: 1500000301 ISM-01 7500000001
0000000300 0.0000008+00 0000me 00000001900 000000015+00
SW01 SWI SW01 SW01 SW01

 

 

3'“:

gserextrx

§§§§§E§§E"

30------"

8
a.

0.000000£+00
SW01
QWEd-(X)
SWE01
QWEM
SWE01
QWE‘I-(XI
SWI
00000001900
SWOE01
000000015400
SNOOE01
QIXXXJOOE-I-(X)
5000000501
QW
SWE01
QWEHX)
SWE01
QW

QWEM o.oooooa=.+00 0.000000£+00 0.000000£+00
SW01 SWI SW01 SW01

0.0000003+00 00000002+00 QW 0.0000005+00
SW801 5.0000002-01 SW01 SW01

0 WEI-(II 0000001900 QW 0000000£+00

SW]
QW
SW01

0000mm

SWE01

SWE01 SWE01
000mm QW 0000000£+00
SW01 SW01 SW01
0.00moos+00 0.0000005+00 0.000000£+00
SW01 SWI SW01
0000000£+00 0000000£+00 0.0000(an
SWOE-Ol SW01 SW01
QW 0000000£+00 QWE-o-Q)
SW01 SW01 SW01
QW 0mm 0.000000£+00
SWE01 SW01 SW01
0000000£+00 0.0000me QWEAD
SW01 SW01 SW01
0mm W 0W

MAL NAMES I
M45) W46) M47) K001'(48)

QW OW QW QW
woman-01 7.5ME01 7500000301 7500000201
o.0000005+00 QWKXI QW QIXXXXXIE‘KXI
SW01 SW01 SWE01 SWE01
0.0000003+00 o.oooooas+oo QW 00000001900
SWI SWOE01 SWE01 SWE01
000010012400 0.000000F.+00 QW 0.000000£+00
SWOI SWOE—Ol SW01 SW01
0.000000s+00 0.00000ua+00 0.0000005+00 0.000000£+00
SW01 SW01 SWE01 SWE01

 

 

I

X00183)

0000000£+00
7M0!
0.000000£+00
SWE01
0000000£+00
SWE01
00000005+00
SWE01
Qme
SWE01
0.000000£+00
SW!
0000000300

~3.: g

£3322t2§

flfl
:9

§§§§§§§§§

8
a.

SWI

SW!
QWEM
5000000501
0.0000001~:+00
SWEOI

50000001501
00000005900

0000000300
5000000301
00000001900
50000001301
0.000000£+00

00000005+00
SWE01
OWN!)

0.000000£+00 QWE-o-(XI QW
SWOE01 SWE01 SW01
QWE-I-(X) 0000000300 0.0000001-:+00
SW01 SWEOI SW01
QW 0.000000£+00 0000000900
SWE01 SW01 SW01
0000000900 0.0000005+00 QW

Figure 32. (continued).

91

§§§§§E§rrssstrsz*

§§E§§E§ssisxtssz.o

 

I 1.000000£+00 11000000900 13000000£+00 1.0000003+00 110000002+00 120000005400

I
I
I
I
I
I

I
I
I
I
I

1.0000002+00 QW 03000001300 0000mm Q mm QW

SIGNAL NAMES I

I
I

K001'(9) KOUI'(10) K0010 1) K00'l'(12)

WEI-Q) 0.000000£+00 Q W 0000000900 0.00000(B+00 QM“)
. 1 .000000£+00 1W 1.0000005+00 101!me
Q 000000£+00 QW Q WEI-(X) 0.0000005+00 QW QWEHX)
Q WEI-(X) QWEAXI SW01 00000005400 QWE‘KXI 00000001900
Q WEAK) QW 1.0000001-:+00 0.0000005+00 00000001900 00000001900
0.000000£+00 QW LWE‘IN 0.00000(£+00 QW 00000001900
Q 000000£+00 omooooswo 1000000900 000000ms+00 QW QWEM
Q WEN!) QW 1.000000£+00 00000005+00 QWE‘I-(XI 0000000900
Q WEI-(X) 0.000000£+00 1000000£+00 0.00000(£+00 QW 0.000000st
0.000000£+00 QWOEHI) 10000001900 QWI-(XI 0.000000£+00 QWE‘HX)
QWEI-Q) QW 1.0000002+00 o.00000a~:+00 QW QWE-I-(XI
0.00000015+00 000000013+00 1.00000a-:+00 QMXIOOE-I-(XI 0.0000001-:+00 QMOOOEAD
QWEW OWE-om LWAXI QWEHX) QWEMXI 000000013400
QWEI-(XI 0.000000£+00 1000000500 QWM 0.0000001=.+00 QW
O.(XXX)OOE+OO 00000005400 LWIE'I-(X) QW OWEN!) QW
QWOOOEHX) 0.000000£+00 1.000000£+00 0.0000001-:+00 o.0000001-:+00 0.0000005+00
00000005400 0000000900 10000003400 QW 0.00000015+00 0.0000001=.+00
QW OWN!) 1W (1W 0. 000000300 QW

I
I

Figure 33: Results of High Gain Ramp Response Neuron With Capacitance.

92

 

TIME I SIGNAL NAMES I
l
(NS) I KOU'KB) 100704) KOUTOS) [00706) KOUTOT) KOUT (18)

a000000£+oo 000000015400 Q W 0W 0.00000a-:+00 0.0000005+00

. LW 1W 10000003400 MIME-HI)
QWEM QW 0. 000000£+00 0.0000005+00 QW 0.0000(nE+00
LWEM LW LWEM 1000000900 LW LUMBER!)
QWEi-(l) QW 00000001900 QW‘KI) QW 0000000900
0.0000me 0.00000015+00 0.000000£+00 QWHXI QW 00000001300

. . QWHX) MIMI-(l) 0.0000001-:+00 QWE-o-(XI
0.000000£+00 QW 0000000£+00 10000001900 QW QWE-I-(XI
OWE-Ml) QW 0000000£+00 1.00000(£+00 QW 0.00000015+00
0.0000001=.+00 QW 00000001900 10000005300 QW 0.000000£+00
0.000000F.+00 0.0000005+00 0000000900 1.00000(£+oo o.oooooor-:+00 00000005400
0.oooooos+oo 0000000£+00 0.0000005+oo 1.0000005+00 00000002400

I
I

I

QW
0.0000001-:+00 0000000500 0000000900 LWEHX) 0.0000005+00 QW
QW 00000008+00 00mm 100000th 0000000£+00 0.000000£+00

. 0W LW 000000015+00 QW
Q WEI-(X) 0W 000000a-:+oo LW omooommo QWE-I-(D

Q WEI-(X) 0.00000015+oo 0000000£+00 LW 00000005300 QW
Q WEI-(XI 0000000£+00 QWd-Q) LW 0000mm QW

§§§§§§§2zraxtrs

I

I

SIGNAL NAMES I

I

KOUT(19) KOUTOO) KOUTOI) KOUTQZ) W) KOUTQA)
0W 0W 0W 0W omooouawo

I
I

QW o.000woz-:+oo Qmm 'Q WEI-(X) LW 0000000500

0000000900 Q W 1000000£+00 QWEHD
QWE-I-(X) Q WEI-(X) QW-I-m Q WEI-(l) IWE-KX)
0.000000£+00 0mm 0.00000th QW
QW 0000000900 0000001500 QW

0.000000%
10000002400 QW
1000000£+00 QM“)

0 QW
4 . . LW LW 10000003400 1000000900
14 I OWE-(X) QW 00000001900 omooamoo QW 00000001a+oo
24 I 1.000000£+00 LW 1000000900 LWW LW LWBM
34 I 0000000£+oo QW 00000005+00 0000000900 QW 0000000900
44 I IWEHX) LW LWEMX) 1.000000£+00 LW LWEW
54 I 0000000300 QW 0000000500 0000000900 0000000900 0.000000£+00
64 I 0.0000005+00 QW QW'HX) 000000aa+00 5W! 0.000000%
74 I QWEKX) QW 0000000300 0000000900 LW 0000000900
84 I QWEM 0.000000£+00 0.000000£+00 o.oooooaa+00 LW‘HD 0.(XXXXDE+Q)
94 I 0.000000£+00 0.000000£+00 00000005400 00000001900 LW Qme
10‘ QWOOOE-I-(XI 00000001900 MIMI-(XI QW LWEKXI 000000013400
3: QWEM 00000003+00 QW Q W 10000001900 0000000300
134
144
154
164

I

Figure 33. (continued).

93

113:3:
(NSII

§§E§§E§22:2:32§:.o_

I

-3

§§E§§§§xrazxtrxzuo

I

KOU'I' (25)

QWE-KX)
1.000000£+00
QWEHXI
l.QXXXXIEHXI
QWEKXI
100000015400
0.0000(an
LWEM
o.0000001~:+00
OWEN!)
00000005400
0.000000£+00
0.0000001=.+00
QW
QWEW
0.0000005+00
0mm
QW

KOUT (31)

I

1000000900

10000005400

0. WBW

Q WEAK)

1.000000£+00

Q 000000£+00

0. MODEM!)
QWW
0.0000001.=.+00
0.000000£+00
0.0000001-:+00

0000000900
0.000000£+00

Karma)

0.00001m+00
1.0000005+00
0.00000013+00
LWEAD
QW
1.000000£+00
QW
LW
0.0000001=.+00
00000001900
10000001900
10000003400
1000000£+00
1W
10000008+00
10000005410
Loooooomoo
1000000300

KOUTGZ)

o.0000001=.+00

KOUTM)

QW
1.0000001=.+00
0.0000005+00
1000000500
0.000000£+00
10000005400
QWEHX)
1.0000005+00
0000000900
QWEM
0.00000015+oo
0.0000005+00
00000012400
0000000900
QWM
0000000900
000mm
0m

KOU'I'(33)

QW
LW

QIXXXXXIBKXI
I 000000900

0000000900

1 .0000005+00

momma

1000000£+00

00000005400
1000000900
00000001900

0000000900
o.00000a~:+00

0000000900
DWI-(X)
QWd-(X)
0000000500
0W

SIGNAL NAMES

SIGNAL NAMES

ICOUT(28)

0W
1W
0000000300
1.0000005+00
00000005400
10000001900
omooooawo
100000t£+00
0000000£+oo
0.000000£+00
0000000900
QW
00000001900
0.00000015+00
0000000900
QW
W
00000003+00

Figure 33. (continued).

94

more»

00000008+00
10000005+00
00000005400
LWOE-o-OO

00000001900
QQXXXDBI-(XI

0.0000me

KOU'I‘GO)

QW
I .(XXJOOOEHX)
0 .00000015+00
1.0000005+00
QWEKX)
1000000900
0.0000005+00
I 000000300
Q 000000500
Q WEI-(X)
0.0000005+00
QWEH'D
0.000000£+00
0000000300
0.0000005+00
QWEW
0.000000£+00
QW

0. WEI-(X)
I 0000001900

0000000900

100000013+00

QWEKX)

LMXXXIE-I-Q)

QWE‘KX)
LWXDEW
QWEA-(XI

50000001501
1 .000000£+00
I .(XXXXJOEHX)
1000000300
LWEHD
1 .000000£+00
1.0000001=.+00

effflifit$§:*o_

'5‘

§§E§§""

________E:2-§-§

EE§£§§2K£K§

iii?

nqsk
(RSI)!

KOU'I'G‘I)

0.000000£+00
LWE-KX)
QWEKX)
1.00000015+00
0000000900
1.0000005+00
Q 000000900
I “WE-+0)
0. 0000001300
I 0000001900
Q WEI-(XI
LWOEHXI
QWE-KX)
0000000900
QWEd-(XI
0.0000002+00
0.0000001.=.+00
QW

awn“)

QW
1.000000£+00
QWEHX)
10000001900
0.00me+00
1 0000001900
0.000000F.+oo
100000015+00
0.000000£+00
1000000900
00000001900
1.0000001=.+00
0.000000£+00
LIXIXXJOEM
00000005400
5000000501
1mm
1W

awn”)

Imam»

00000003+00
LWW
QW
I 00000054»

KOUT(39)

Q W
I 0000005400
0 000000900
1000000£+00
o.00000m~:+00
1.000000£+00
0.000000£+0o
1.000000£+00
000mm
LWM
0.000000£+00
1000000500
0000000300
o.00000aa+00
0.00000as+00
0.00000t15400
0000000900
0W

awn”)

Q W
I 000000le
Q WEI-(XI
LWEM
0000000£+00
1000mm
QWEKXI
1W
0000000540
1000mm
0000000900
1000mm
0.00000aa+00
Ham-K!)
0.0000005+00
(1000000300
0.000000I.=.+oo
0000mm

SIGNALNAMES

SIGNAL NAMES

KOUTGO)

KOUTGQ

0W
1000000900
00000001900
1.00000a=.+00
0.00000(B+00
I 000000300
0. 000000500
1 0000005+00
0. W
LIXXXXXE'KX)
0000000£+00
LW
QWEd-Q)
LW
QWEd-(X)
QWEM
QW
QW

Figure 33. (continued).

95

zany»

mmmn)

0000000300
LWM
0.000000£+00
LWEM
QW
1.000000£+00

ammo
100mm
QW
1.000000£+00
00000005400
1.0000005+00
00000003400
10000001900

KOUT(42)

QW
1.000000£+oo
0.(XXXXX)E+(XI
1.000000£+oo
0.000000£+00
1.000000£+00
0000000300
1000000900
00000001900
1.0000005+00
0000000900
1000000900
Q ooooooa+00
Q 0000005+00
0. W
00000001900
o.0000001~:+00
QW

awn“)

QW
LW
0.000000£+00
1.0000005+00
0.0000005+00
10000005+00
0.000000£+00
I .0000001=.+00
QWKX)
10000001900
0.000000£+oo
l .0000001-:+00
000000013400
I .WEHX)
QWEI-Q)
QWE-I-Q)
0000000900
QW

 

mm .L SIGNAL NAMES 4
018') I XOUTO) KOUI‘Q) KOU'I‘G) [007(4) KOU'IIS) KOUT(6)

0 I 0.000000£+00 00000112400 QW 00000001900 QW 0.00000ma+00
4 I Sim-OI WI Sim-OI Sim-01 SSW-01 8W0]
0W 0W QW QWE‘IN QW
LWOEM o.oooooua+00 0W 00000005400 00000001900 QWEi-(X)
10000001900 0W 0W 0.0000005+00 00000005+00 QW
1.00000054-00 0.00000tn400 OW QW 0.0000001=.+00 QW
1.0000001-:+00 QWI-(X) QW 03000001900 0.00000015+00 QW
LOWE-rm 0W QW o.ooooooa+00 0000000900 QW
LWOEHX) omooormoo QW QW OWE QW
1.000000£+00 0.00000aa+00 o.ooooooaoo QW omoooom QW
LW 0.00000aa+00 QW QW DWI-(X) QW
1000000900 QW 0 0000005+00 0W 0.00000mwo QW
1000000st omoooaswo 0 0000003400 0W QWM QW

 

1000000500 QW o.oooooas+00 0W 0000000900 0000000300
IWEW QW 0.00000m00 0000000900 01100000900 omooooaoo
10000001900 0.00000012+00 00000005300 QWEHX) 0.00000ua+00 QW-o-(X)
1.000000£+00 QW 000mm QW 0000000900 09000005400
10000005410 QW 00000005400 0000000900 00000005400 QW
INS-0m QW omoooaswo 00000005+00 omoootmoo 0W

§§§§§§§E§E§EEEEEfisssssssss
I
I
I

 

 

TIME' I— SIGNAL NAMES I
(NS') I KOUTG) K0011!) K0079) KOUTOO) KOUTO I) KOU'KIZ)

0 I 0.0000005+00 0000000300 QW 09000005400 QM“ QM“)
4 I QWE-OI SWOI SW-OI SWE-OI W01 W01

I4 I 00000001900 0000000540 00000005400 0000000900 QW 00000001900
24 I IJOOQXIE-OI 09000005400 5000000301 0000000900 1500000501 MIME-HI)
34 I 0.000000£+00 0000000900 1W0] QW 0.000000£+00 QW
44 I amount-300 0.0000001~:+00 BW-m 0.00000015+oo 0.000000£+00 QWEKD
54 I 00000001900 0.0(1000015+00 Sim-OI 0.000000% 0.0000ws+00 QW
64 I 000000015+oo 0.000000% Sim-OI QWOEHX) 0000000900 QWW
74 I 0.0000001.-:+00 QW SSW-01 QW 0.0000005+00 QW
34 I 0000000£+00 QW 8501115401 QW 0000mm QW
94 I QWEi-(X) onmoom-oo SSW-01 QW 0000000900 00000001900
104 I QWEM 00000003+oo 8W0! 0.00000013+00 QW 00000005100
114 I QW 0.0000005+00 8W0! 00000001900 QW OWEN!)
g: : o.oooooox-:+00 ammo asooooma-m QW-KXI 000000012400 0W

QW 0mm 8W0! QWEM QW omooooawo

Figure 34: Results of Low Gain Sigmoid Response Neuron With Capacitance.

96

§§§§§§§

93
a

EEEEE

0.0000005+00
QWEHXI
Q WEI-(XI
Q 0000005+00
Q (XXXIOOE-o-(XI
Q WEAK)

Q WEI-(I)
0.000000£+00
Q WEI-(X)
Q MODEM
aoooooomoo
QW

WOI
SSW-01
8W0]
8W0!
BW-m
SW-OI
SW-Ol
woman-01
SW01
momma-01
Sim-01
Sim-01
W01

97

Figure 34. (continued).

 

 

 

 

TIME I SIGNAL NAMES I
I
(NS) I KOUT(13) KOU'I'(I4) KOUTOS) KOU'KI6) KOUTO‘I) KOUTOS)
I
0 I QW 00mm QW 00000003+00 0.0000002+00 Qm-I-(X)
4 I W01 “000005-01 Sm01 SW01 SW01 SW01
14 I 0000000900 Q WEI-(XI QWEi-(XI 0.00000(E+00 QW 00000002410
74 I SWBOI SW01 50000001301 SW01 SW01 5000000201
34 I Q 0000005400 QWEI-(D DWI-(XI 0.000000£+00 QW QWEI-Q)
44 I IWE01 IWEOI 1W1 S.WE01 SW01 1W0]
54 I 0 OWE-(X) QW 000000013400 0.00000aa+00 QW Q WEI-(X)
64 I QWE‘HX) LSQXXIOEOI 0.0000005+oo SW] IWEOI IWEOI
74 I QWE+00 QWE-o-Q) QWEM 1W0! Q WEI-(X) QWEI-(II
84 I 0000000900 QW O .0000005+00 SW01 LSWEOI QW
94 I QWE‘HXI 0.000000£+00 Q 000000900 5000000501 OWE-0m QW
104 I QWE-I-(X) QWEHXI Q 000000900 SW01 0.0000005+00 0.00000015+00
114 I QWEW QWEKXI 0 .W SW01 QWEM QWE-I-(X)
124 I 0.0000001=.+oo 00000001900 00000005400 SW01 0.000000% 0.0000005+00
134 I 0000000900 0000000900 00000001900 SW01 QW QW
144 I 0.0000001-:+00 QW-I-Q) 00000005400 W01 QWEHX) QW
154 I QOQIOOOE-I-Q) 0.000000£+00 0000000500 SW01 0.0000002+00 0.0000005+00
164 I QQXXIOOM 0.0000001~:+oo 0.000me SW01 0.0000005+00 QW
174 I QWEM QWEKXI 0.00000aa+00 SW01 o.00000015+00 0.0000001=.+00
IS4 I QWEM QWI-(X) QWEM SW01 0.000000£+00 00000001900
194 I 0000000900 Q WEI-(XI 0.000000£+00 SW1 QW 00000005400
20! I QWOEM 0 .(XXXDOEHXI 0000000300 SW01 QW QWW
214 I 0.000000£+00 Q WEI-(X) 00000001900 SW01 QWEHD QWEM
274 I 0.0000005+00 0000000£+00 0 .000000£+00 SW01 o.m00001.~:+00 0000000900
234 I 0.0000005+00 0.0000008+00 Q 000000£+00 SW01 QW 0.0000005+00
244 I QWEM 00000005400 0 .0000005+00 SW01 QW QWEM
254 I 0.0000005+00 0000000900 oncoootmoo W1 00000001900 OWE-(XI
264 I 0.000000£+00 0.00000aa+00 0mm SSW-01 W QW
TIME I SIGNAL NAMES I
I
(Nsl) I KOUI(19) KOUIIZO) K0010!) K0073?) K0011”) ,KOUTQA)
0 I 0000000£+00 0W4“) Q W 0W 0.0mm a0000005+00
4 I SWEOI SW01 SW01 SW50! SWEOI SW01
I4 I OWE-(XI 0.0000005+00 Q 0000005+00 omoooaa+oo Q 0000005+00 Q OWE-(X)
24 I SW01 S.WOEOI 5. W01 S W01 S 000000501 5 .OQXXXJEOI
34 I 00000001900 00000001900 Q WEI-(XI QWI-(X) 0000000500 00000001900
44 I 5.0QJOQJE01 S.WE01 S.(XXXXIOEOI S.QXXI(XIEOI 5000000501 S.QXXXXIEOI
S4 I OWEN!) 00000001900 0.000000£+00 QWXXEHXI QW 0.000000£+00
64 I SW01 SIXXXDOEOI 10000001501 SW01 SIXXXXDE01 S.QXXX)OEOI
74 I 0000000500 QW 0000000£+00 QWd-(X) QWEM QMOOQIE'HX)
S4 I SWE-OI SW01 S. W01 5410000301 S .WEOI S .WOEOI
94 I OWEW QW Q 0000005200 Q 000000900 00000005400 0 000000900
104 I LSIXXXIOE01 1W0] SW01 SW1 S 000000201 5 000000501
114 I 0.000000£+00 0.000000£+00 Q oooommoo Q W Q 000000£+00 Q WEAK)
124 I 0000000300 Im0l mom-01 1111111301 SW01 15000001301
134 I 0.000000£+00 ammo omooormoo QM“) SW01 0W

Figure 34. (continued).

98

§§§§§§E

B
;

§§§¥§_-_-_-__

I

I

:gtraar:r§:.o

”§§§§§§E§§""

a.
fi

§§§?§___

0.000000£+oo
0.0000001=.+00
QOQJOOOE-I-(XI
QWEHXI
Q WEI-(X)
Q 0000001900
Q 000000£+00
0.0000001=.+00
QWEHD
QWOOOEd-(X)
QQXXIOOE-o-(XI
00000001900

QW

KOUTOS)

0.000000£+00
SWEO!
0.oooooos+oo
SW01
0.000000£+00
SWO!
QWEHXI
SWO!
00000001900
SWE01
0.0000005+00
S.(XI)OOOE01
QOIXXIOOE-o-OO
SW01
QIXXXIOOEM
1.5MOOE01
QQXXIOOE+OO
QWOOOEKI)
QWOOOEW
0000000900
QWOOOE-I-(XI
QWOOOB-o-(XI
00000001900
QWOOOEM
QWOOOE-KX)
QWOOOEHX)
QWEAX)
0W

00000003+00
0.000000P.+00
QWEW
00000005+00
Q 0000003+00
0000000£+00
Q WEN!)
0.0000(an
QWE‘o-(XI
00000005+00
0000000300

0.000000£+00
000000aa+00
0.000000£+00
0.00000m+00
0.000000£+00
000000t£+00
0.0000005+00
onooootmoo
0000000300
0000000300
QWI-Q)
00000001900

W
QW
00000005400
QWE-I-(XI
0.000000E+00
QOOQIOOEHX)
QW
QWEM
000000015200
QWEHX)
0.0000001-:+00
00000001900

SW01
SW01
SW80!
SW50!
SW50!
SSOOOQIEOI
SW01
SW01
SW50!
SW50!
SW50!
SW50!

QW
ommooE+00
Q OOQIOOEHXI
0 .(XXXXIOE-O-OO
0 000000500
Q MOEHX)
QOOQJOOEM
QOOQXIOEAX)
QWEW
QWEM
0.0000001a+00
QWEAX)

0.0000005+00

KOU'I‘(26)

Q 000000£+00
SWO!
Q WEI-III
SWO!
QW
5000000501
00000001900
SW01
QW
SWE01
QW
SW01

0.000000£+00

SW01

o.0000001=.=.+00

SW01
LSQXIIOE0!
SW50!
SW01
SW01
SW01
SW01
Sm0l
SW01
SW01
SW50!
SW50!
SW01

omoootmoo QW Sm01 QW

 

SIGNAL NAMES I

KOU'I'(2‘7) KW) KOUT(29) KOUT (30)

QW OWE-04X) 0000mm 00000005400
SW01 SW01 SW01 SW30!

0000000£+00 00000001900 Q WEI-(X) 00000001900
SW01 S.QXXXX)E01 1W0! SW01

0000000£+00
SW01
0.0000001-:+00
5000000501
000000015+00
S. 000000201
Q 000000£+00
S .WEOI
Q 000000900
SW01
0000000900
IWEO!

5. W501 S 0600001501 5 .WOBOI
Q 000000900 Q 000000900 Q WEI-(X)
SIXXXJOOE0! SW01 SW01
0000000300 0.00000013+00 0000000300
S.(XXXIOOBOI SW! SW01
QOIXXXXIE-I-(X) 0.000000£+00 QWEW
lSQXDOE0! 1.S(IIIXDE01 1500000501
0.0000005+00 0000000900 QOIXXXIOEAX)
0.000000£+oo 00000001900 00000001900
QWEI-Q) 0000000900 QWEAX)
0000000900 QWEM QWEAX)
0.000000£+oo QW 0.0(XXXJOE-I-(X)
00000001900 QWEHXI QWEHX)
0.000000£+00

 

QWEIOO
QOOOOQIE‘I-(XI
0000000300
OWEN!)
0000000500

Figure 34. (continued).

99

0.000000£+00
QWEM
QWEM
0.000000£+00
00000001900
QW

01!!!me
QWE-I-OO
00000001900
QOIXXIOOEM
0.(XXXX)0E+IXI
QWEI-OO

SIGNAL NAMES
KOUT (33)

 

1?

I KOUTO 1) i KOUTOZ)

000mm
S.m01

E?

woman-01
SW!
SW01
SW!

"7%..."

i6£§§§§§i§§i§§§§§rrrzr:22::2_3

9

o.oooooaa+oo

' SIGNAL NAMES
mums) mums»

onoooooaoo
SW01
o.oooooo£+oo
SWO!
QW
S.WE01
QWE-I-m
1W0!

 

1-

KOUTG‘T)

é-a

W0!
10000001301
S. m0!
S. W01
S .ooooooa-on

rrrzrtrxz‘o
é

S .ooooooa-m

SW01 SW01

§§E§
u
E
S:

Figure 34. (continued).

100

[(00164)

QW OWN o.oooooua+oo 00000005400
SW01

QWEd-(X) o.oooooaa+oo o.oooooo£+oo o.oooooo:-:+oo
S.WEO!
00000001900 onoooooewo aooooooaoo QWEKX)
SWO!
o.ooooooa+oo onooootmoo QWEAX) o.oooooos+oo
SW50!

00000005400 o.oooooaa+oo QWW QWBM
S .WDOEOI

KOUTGS)

Sm0l W01

SWE01

1W0!

S .ooooooe-ox
S.WE01
SW!

SW01
QW

0900000300
o.ooooool-:+oo
QW
o.ooooooI-:+oo
QW
QW
o.ooooooa+oo
aooooooa+oo
ammo

KOUTGO) XOUTG!)

QW 0W omoootmoo
SW01

SW01
5 .WE01
S .WE01
S .mB-O! S .WE01
S .WE01
S .m01

SW30!

1W0!
S.(XXXXX)E01

SW01
QW Q WED-(X) o.oooooo£+oo
SW01
QWEHD o.ooooooE+oo QW
SW01
o.oooooor~:+oo QWEd-(X) o.ooooom=.+oo
SW01

S.WE01
S.WE01
SW!

o.ooooooe+oo
1W0!
0.(XXXXX)E+(X)
o.oooooo£+oo
o.oooooo£+oo
o.ooooooa+oo
09000005400
o.ooooooa+oo
OMAN)
onooooomoo

S .WOE01

KOUTGG)

1000000501
1W0!

09000001900

SW01

5111111501
S.(XXXXX)E0!
SW01
S.S(XXXJOE01
”000005-01
8.5mE0!
SW01
SW01

KOUTGZ)

QW

SW01

o.ooooooz+oo Q ooooooa+oo Q W 0000000900
S. W01
Q 0000005qu QWM QW QWE-o-m
S .WI
uoooooomoo omooooewo o.ooooooa+oo 0900000900
1000000201

Q oooooomoo omoooos+oo o.ooooooa+oo 09000005400
S .QXXXJOE01
aooooooaoo o.oooooaa+oo o.ooomoa+oo o.oooooor~:+oo
S .oooooos-m

o.oooooos+oo Q ooooooE+oo QW o.ooooooE+oo 0300000300
SW01
0W 090000th QW 0mm 0300000300

S .WEO!
SW01

5 .W01

5.000an01

 

 

144 I 1000000501 SW01 SW01 SW01 S .W! SW01
154 I o.ooooooa+oo Q 000000300 o.oooooaa+oo QW Q WEI-(X) Q W
164 I 5110000501 S .oooooua-on SW01 SW01 S .WEO! S .oooooos-ox
0. “0000510) 090000013400 09000005400 0. W 0. WEN!) 0. oooooomoo
184 I S. (111)00501 S .m01 momma-01 S .ooooomm S. W01 S .oooooaa-on
194 I QWOOOEM Q ooooooE+oo o.ooooooe+oo 0. W 0 .ooooooa+oo 0. W
204 I S.(XXXXIOE01 S .ooooooe-m 1W0! S .WOE01 1. 500000801 1W0!
214 I 09000001900 o.ooooooa+oo QWM 0. ooooooa+oo 0. WEI-(XI 0. W
224 I 15WE01 1W0! 1W0! SW01 QW 15MB01
234 I 0. W o.ooowos+oo QMM S. cocoons-01 0.00000054-00 QWEHX)
244 I 0. WEI-(X) QW QWHX) 5W0! QWEW o.ooooooa+oo
254 l 0. ooooooswo 09000001300 omooomwo SW01 o.ooooooE+oo o.oo:noo£+oo
$4 I 03000002400 0mm ammo SW01 QW ooooooomoo
TIME I SIGNAL NAMES I
I
018') I KOU!‘(43) K0011“) KOUTGS) W46) KOUT(41) KOU'1‘(4S)
0 I QWEW 09000001900 0. W 0W 0W 0. ooooooaroo
4 I W01 S.m0! SW01 SW01 SW01 SW01
14 I ooooooo£+oo o.ooooooa+oo 0mm omooocmoo QW o.oooooos+oo
24 I ' SQXXXXIE01 SW01 1W0! 501010501 SW01 5 .WE01
34 I OWE-(X) QW 0W 0. oooooaa+oo 0. 0000mm 0 .oooooomoo
44 I _ SW80! SW01 SW01 S W! SW01 SW01
S4 I o.oooooox-:+oo QWHXI 09000005400 OHM-Kl) QW QW‘IN
64 I 1000000501 1W0! SW01 10000001301 S.WE01 S.WE01
74 I omooooaoo QW o.ooooooa+oo omooooawo QW 09000002400
S4 I SW80! 1mm 1W0! SW01 SW01 SW01
94 I QWE'o-m QWHXI 09000008400 o.ooooous+oo QW o.oooooos+oo
104 I 5111000501 SW01 SW01 SW01 1000000501 SW01
114 I QWEAX) OWN!) omoootmoo QWEHX) o.ooooooa+oo o.oooooos+oo
124 I S.QX)0(X)E0! SW01 SW01 SW01 1W! SW01
134 I QWEM 0W 0. ooooocmoo QW 0W QW
144 I 1000000201 SW01 S .m01 S.WE01 SW01 SW01
154 I 0111100544.!) ammo omooooewo QW omoooomoo QW
164 I S.(XXXX)OE0! SW01 5 .WE01 SWOE-O! SW! SW01
174 I QWEM ooooooomoo 0. ooooooa+oo QW o.oooomE+oo QWKX)
184 l S.WE01 smoooaa-m S .m01 SW01 1000000301 SW01
194 I QWEM 0. WEI-(X) QW‘KX) 0.000000134-00 0.(XXXXXIE+(XI QQXXXXIEM
204 I S .WOE01 51100000341 10000002411 SW01 SW! SW01
214 I 0. MOE-14X) Q 0000005400 0. ooooooa+oo QWE-I-(X) o.ooooooB+oo QQXXXXIEM
224 I 10000001501 S .ooooooa-o: S .(XXXXDEO! SW01 SWE01 SW01
234 I 0. oooooos+oo QWE-HX) 0. 0000001900 o.oooooos+oo QWEM o.ooooooa+oo
244 l SQXXXIOE01 QWM 0. WEI-(X) 0. W40) 0. oooooos+oo QWE-I-(XI
254 I Loooooos+oo 09000001900 omooooaoo 0. W o.oooooo£+oo 0300000900
264 I 1mm 0mm 0mm 0mm omooooawo QWEHXI

Figure 34. (continued).

101

:rrtrcrz

§§E§§E§33

I
mSI) I KOUTO)

Iii?

KOU'RZ)

0W
Loooooos+oo
QWEW
ooooooamo
ooooooomoo
0. W44!)
0. ooooomwo
ammo
QW
oooooooewo
omooooewo
uoooooomoo
uooooooE+oo
uooooooewo
QW
o.ooooooa+oo
o.oooooos+oo
QW

KOU!‘(3)

olooooocmoo

SIGNAL NAMES
KOUT (4)

omooooa+oo
mooooomoo
QWEKX)
QW
o.oooooor-:+oo

obwwoem

KOUT(S)

K001“)

QQXXXXIE‘I-(X) 0W

LWM

o.ooooooa+oo
o.oooooon+oo
o.oooooor-:+oo
QWEI-m
00000001900
QWEKX)
o.oooooor.+oo
ooooooomoo
QWEI-(X)

QWM
o.ooooooE+oo
0W4“)
o.ooooox£+oo
o.oooooos+oo
o.oooooos+oo
ammo

1 .omoooam

7%

SIGNAL NAMES I
KOUTG)

S

KOU!‘ (10) KOUTO 1)

r::zx:rc::3_§

§§E§§E§

uoooooomoo
Loooooomoo
omoooom-oo
QWEM
QWE'I-m
o.oooooor-:+oo
OWEN!)
OWEN!)
OWEN!)
03000001900
OWEN!)
QWE'I-m
QWE-I-(X)
0. WEI-(X)
0. (100001544!)

ooooooomoo

o.oooooos+oo
LW
o.oooooos+oo
S.QXXXX)E01
1.0000(an
Loooooomoo
LW-KX)
wooooos+oo
1.000000st
140000005410
LWE‘KX)
Looooooewo
Loooooomoo
Looooowwo
moooooa+oo
1 .ooooooaoo
moooooswo
LW

o.ooooooz+oo
hm
09000001900
0. 000000900
0. ooooooawo
0. ooooomwo
omooooa+oo
QWM
04000000900
QWKX)

ammo
moooooewo
QWEM
QWE-I-m
o.omoooe+oo
QW
00000001900
QW
001me
QW
uooooooaoo
QWEM
o.oooooo£+oo
o.ooooooa+oo
0.000000%»
o.ooooooE+oo
OWEN!)
onooooomoo

KOUT(12)

QW
LWEM
OWE-04X)
QW-I-(X)
QWEd-(X)
o.ooooooE+oo
o.ooooooI-:+oo
0. oooooos+oo
0. (“KNEW
0mm
QWE-om
QWEHX)
QWE‘KD
QWEHXI
o.ooooooa+oo
0. WEI-IX)
0. MBA!)
QW

Figure 35: Results of High Gain Sigmoid Response Neuron With Capacitance.

102

 

TIME k SIGNAL NAMES I
I
W) I KOUTOS) KOUT(14) KOUTOS) KOUTOG) [(001117) KOUTOS)

01100000900 0. W 0W QW QW
LMW 1 .W LW mooooa-mo LWE-o-(X)
QW 0. mm o.oooooas+oo QW QWE-I-(XI
LW woooooawo Loooooaawo LW Looooooem
QW omoooonwo 09000005400 QW omoooomoo
. o.ooooooa+oo 0. ooooooa+oo QW omoooomoo
QM“) omooooaoo lmooooawo QW o.ooooooa+oo

o.ooooooa+oo QWM 1 .ooooooawo QWE‘KXI o.oooooor~:+oo

QW 09000002400 1W QWHII o.oooooos+oo

QW omoooomoo Loooooa-zwo 0.00000054-00 QWEM
040000005400 o.ooooooa+oo omoooomoo LWIEi-(XI QWEHX) 0900000300
QW oooooooswo QW LW o.oooooo£+oo QWEM
QW omoooomoo o.oooooaa+oo LW o.oooooos+oo QW
o.oooooos+oo 01000000900 o.ooooo(E+oo LW oooooooewo 0.00000054-00
. momma LW o.ooooooa+oo 0.0000005+oo
0300000500 onoooooa+oo QWM LW o.oooooos+oo 0300000300
mmoomoo omoooomoo o.oooooaa+oo LW 0000000300 09000001900
W omoooomoo ooooooomoo LW omoooomoo o.ooooooz+oo

éééééééééé
I

§§E§§E§:::r::::::g_

I

<3;

SIGNAL NAMES I

~3.

KOUT(19) KOUTCN) KOU'!‘(21) Km) KOU'I‘(23) KOUTQ4)

QW 0W moments-+00 0W 0mm QW
1000000900 1mm LW 1W Loooooomoo LW
QWEM QW 0900000900 QW‘HX) 0.00000054-00 o.ooooool~:+oo
nmooooewo 1.oooooo£+oo MIME-IQ) 19000001900 LW 1 WEI-(X)

OWE-Kl) QW o.ooooooa+oo omooooawo QW 02000000300
moooooa+oo l.QXmOE-o-Q) LWHX) moooooswo Lamoooem

0

4

14

24

34

44 Looooooe+oo
S4 QWE+00 QW QW o.oooooaa+oo 0. W o.oooooo£+oo
64 o.ooooooa+oo QW QW o.ooooom+oo S. W0! o.oooooos+oo
74 omoooom-oo QW 0. WEI-(X) o.oooooaa+oo 1 .ooooooewo 01!!!!an
S4 0000000900 0.00000054-00 0. WEI-Q) QIXXXIIXEJ-(II 1.00me 0000000900
94 o.oooooos+oo QW 001100844!) o.ooooom=.+oo LWE-I-(X) o.ooooooI-:+oo
104 I QW 0000000900 0. oooooamo o.oooooos+oo 1.0000me 0.00000054-00
114 I o.oooooor-:+oo QQXXXIOE'KXI o.ooooooE+oo QW 1.ooooooa+oo 01111110544!)
124 I 0.0000005+oo 0 .ooooooewo QWM QW Loooooos+oo QW
134 I o.ooooooa+oo 0. 0000001900 o.oooooos+oo QW LWE-I-(XI QW
144 I 001000514!) o.ooooooe+oo o.oooooaa+oo 0.000()o01=.+oo 1.ooooooE+oo QWEI-(XI
:2 : QW ooooooomoo oooooomwo QW 10000002400 QW

oooooooswo QW 0W QW 1.oooooo£+oo QW

Figure 35. (continued).

103

I -——

 

"3-;

eugggggrrxertxnzte

Hide-Ionian...—
--’-—-

smmunums g

I xmnwn mmNM) Immmn

o.oooooo£+oo
LWOEM
QWEM
1.oooooos+oo
QIXJOOQIEAX)
l.QXXJQIEd-(XI
QWE-I-(X)
l.QXIOMEm
QWEKD
QWE'HX)
omoooosm
QWE-o-m
QWOEHX)
0. WEI-(X)
0. 0000mm
0. WEI-(X)
o.oooooor-:+oo
QWE'I-N

0. W
1 .ooooooem
0. 0000001900
1 .oooooomoo
0. 000000500
1 “WEI-(X)
0. 0000001900
momma-00
QWEW
QWEM
09000001900
QWEM
00000005400
00000002400
QW
o.oooooos+oo
o.ooooooe+oo
QWW

0 .oooooomoo
10000001900
QWE-KX)
LWE-I-m
0. oooooos+oo
1 “HIDE-KI)
0. oooooos+oo
0000000900
0000000900
QWEHI)
o.oooooos+oo

QWE‘I-m
QWEM
0. WEI-(X)

0W 0 mm

03000000900

1-

42

autrgzgttrflrtrfizeo

Multiple-ope...
---—

KOUI'G 1)

ooooooo£+oo
Loooooos+oo
000000012400
LWE-I-m
QWE-I-(X)
LWEMX)
o.oooooo£+oo
”NINE-0m
o.oooooos+oo
l.QXXXXIE-o-m
o.oooooos+oo
QWOOOEKD
o.oooooon+oo
QQXXIO0E+M
o.ooooooa+oo
0011000544!)
QIXXIOOOEHX)
QIXIXIOOEKX)

KOUTGZ)

OM44!)
Loooooo£+oo
QWEM
1 .ooooooswo

KOUI'GS)

0. W
1 .ooooooawo
0 .ooooooswo
Loooooomoo
QWEKXI
1 .oooooomoo
Q WEI-(XI
1 .oooooomoo
0 “WEI-(X)
l.QXXXXJEm
QWEM
QM“
oooooooewo
o.oooooos+oo
QWXXEm
0900000300
o.oooom!.=.+oo
0W

mmnu)

omoooomoo
LW

QWHX) '

Looooooem
QIXXXXXEHX)
1110001544!)
QWW
1.ooooooE+oo
QWM
11!me
00000005400
QW
QW
QIXXXIOOEHII
o.oooooo£+oo
o.ooooooa+oo
QWEHD
QW

Figure 35. (continued).

104

XOUTGS)

0W4“)
10000005410
0.00000054-00
l.QIXXXIE-om
QWEHD
LW
o.oooooos+oo
1 .ooooooswo
Q WEI-(X)
1 .ooooooawo
o.ooooooa+oo
o.ooooooa+oo
o.ooooooa+oo
QWE‘KX)
QWEM
Q WEI-(X)
omoooomoo
omoooo£+oo

$IGNAL NAMES g

KOUTOG)

QW
Looooooewo
anemone-+00
Looooooa+oo
o.oooooos+oo
LWE-I-(X)
o.oooooos+oo
1 900000900
0. WEI-(XI
1 .ooooooem
090000015400
S. oooooos-on
Loooooos+oo
1.oooooos+oo
l.QXXXXJE-KD
1.ooooooe+oo
1 .ooooooE+oo
LW

Mk

 

 

 

 

Figure 35. (continued).

105

5101141. NAMES 1

1

0451 1101111371 50111130) 110111139) Karma) 1101:1141) 1101:1142)

1

01 00000005400 00000015400 0. 0000005400 0000005400 0000005400 00000005400
41 1.0000005400 1.0000005400 1.000005400 10000005400 1.0000015400 1.0000005400
141 0.0000005400 0.0000005400 0. 0000005400 0.000005400 0.0000005400 0.0000005400
:41 1.0000005400 1000005400 1.0000005400 1.000005400 1.0000005400 1.0000005400
341 00000005400 0.0555400 0.5005545 0.5000545 0.0055545 0.5005545
441 1.5500545 1.0055400 1.0000005400 1.000005400 10000005400 1.5005545
:41 0.0000005400 0.0000005400 0.0000005400 0000005400 0.0000005400 0.0000005400
641 1.5500545 1.00me400 1000005400 1.000005400 1.5555400 1.5000545
141 0.0000005400 00000005400 0.5500545 0.0555400 0.000005400 0.000005400
1141 1.0000005400 1.5005545 1.0005545 1.0050545 10000005400 10000005400
1141 00000005400 00000005400 0. 0000005400 0.000005400 00000005400 0.0000005400
1041 10000005400 10000005400 1 .000005400 1.000005400 10000005400 1.000005400
1141 0.0550545 00000005400 0000005400 0.0000005400 00000005400 00000005400
1241 00000005400 00000005400 0000005400 5000000501 00000005400 00000005400
1341 00000005400 0.0055545 11. 000005400 10000005400 00000005400 00000005400
1441 00000005400 0.0000005400 0. 000005400 1.0000005400 0.0000005400 00000005400
1541 00000005400 0.000005400 0000005400 10000005400 00000005400 00000005400
1641 00000015400 0000005400 0.0000015400 1.0000015400 00000015400 0.0000005400
1111451 SIGNAL 514145: e

1
(14.4.11 110111143) tau-1144) now-(45) zoo-1146) 151111141) 110111140)

01 00000005400 00000005400 0. 000005400 0. 0000005400 0000005400 0.0000005400
41 10000005400 10000005400 10000005400 10000005400 10000015400 10000005400
141 00000005400 00000005400 0. 0000005400 0 .000005400 0.0000005400 00000005400
241 1.0000005400 10000005400 10000005400 1.000005400 1.0000005400 10000005400
341 00000005400 00000005400 0.0000005400 0.000005400 00000005400 00000005400
441 1.0000005400 10000005400 10000005400 10000005400 10000005400 10000005400
541 00000005400 00000005400 00000005400 0.000005400 00000005400 00000005400
1541 10000005400 10000005400 10000005400 1.000005400 10000005400 10000005400
141 00000005400 00000005400 00000005400 0.000005400 00000005400 00000005400
1141 10000005400 10000005400 10000005400 1.000005400 1.0000005400 1.0000005400
941 00000005400 00000005400 00000005400 0.000005400 0. 0000005400 0.0000005400
1041 10000005400 10000005400 1.000005400 10000005400 10000005400 10000005400
1141 00000005400 00000005400 00000005400 00000005400 00000005400 00000005400
1241 10000005400 10000005400 1.000005400 10000005400 10000005400 10000005400
1341 0. 0000005400 00000005400 0.000005400 0.0000005400 00000005400 00000005400
1441 5 000000501 0000005400 00000005400 00000005400 00000005400 00000005400
1541 10000005400 00000005400 00000015400 00000005400 00000005400 0.0000005400
11541 10000005400 00000005400 0.000005400 00000005400 00000005400 00000005400

 

 

. QW QM
4 I 1000005400 1W LW 10000005400 10000005400 LWE-tg
14 I 10000005400 0000005400 00000005400 0.0000005400 QWEM QWOOOE'I-(XI
24 I 1 WEI-(X) 0 000005400 QWEHI) MIME-14X) 0 WEI-(XI 00000005400
34 I 1 WEI-(X) 0 000005400 QW QW 0 WEI-(XI 00000005400
44 I 1 OOQXIOBHX) 0 W 00000005400 QW 0 0000005400 00000005400
S4 I 1 WEi-OO 0 000005400 QW QW OWEN!) QWEHD
64 I 1 WEI-00 0 000005400 QW QW 0 WEI-Q) 00000005400
74 I 1 WEI-(X) 0 000005400 QWEHX) 00000005400 0 WEN!) 00000005400
S4 I 1 0000005400 0 000005400 QW 00000005400 0 0000005400 QW
94 I 1 MODEM 0 000005400 QW QNIXXIOE-HXJ 0 WK!) QWE‘KX)
104 I 1 0000005400 0 0000005400 QWM QWEHX) 0 000005400 00000005400
114 I 1 MEN» 0 W00 00000005400 00000005400 00000015400 QWEM
124 I 1 11(1) 0 W 0000005400 00000005400 0 000005400 QWE-HI)
134 I 10000005400 QW 0000005400 0W 0000005400 00000005400
115:8 !—~ WSIGNAL NAMES '

 

-3

I KOUTCI) KOU'1'(S) KOU'I'(9) KOUTOO) KOUT(1 1) KOUTOZ)

25552532254:

I
1

iii?

Figure 36: Results of Three Stage Interconnection Pattern and Binary Neuron.

106

MI

 

 

 

 

Figure 36. (continued).

107

SIGNAL NAMES I
I
(NS) I K007'(13) KOU7(14) K00'1'(1S) K007(16) K00'l‘(17) K00'1‘(1S)
I
0 I QW 00000005400 QW 0W 0.000005400 W
4 I 11110005140 1.0000005400 LW “RIDGE-(XI 10000015400 1.5500545
14 I 0.0055545 0.5005545 00000005400 QWE'KXI QW 0.0000005400
24 I 1010001544!) LW 00000005400 1.000005400 QW 0.5555400
34 I OWEN!) QW 0.0000005400 0. 000005400 QWE-I-(XI QWEKXI
44 I 00000005400 00100005400 0W 1000005400 QW 0.5005545
54 I 0.0000005400 QW QWEKX) 1000005400 QW 0.0000005400
64 I QWEKX) QW 0.0000005400 1000005400 QW 0.5500545
74 I 0.0000005400 QW QW 1W QW 0.0000005400
S4 I 0. WEI-Cl) QWH'D 0.0000005400 1.000005400 QW QWE-KX)
94 I 00000005400 00000005400 00000005400 1.000005400 QW 00000005400
104 I Q 0000005400 0.5055400 0.5505400 LW QQXXXIIE-I-(XI QW
114 I 0.0000005400 0.0000005400 0.0005545 1.5500545 QWEAXI QW-HXI
1214 I QW 0W 0000005400 LW 0M4“) QW
134 I W 0W W LW 00000005400 00000005400
TIME I SIGNAL NAMES I
I
(NS') I K007(19) KOUTW) K0070!) K007(22) K007(23) K00704)
0 I QW 0.055545 0. W 0.0005545 0.5000545 0W
4 I 1.5500545 1.5555400 1 AIME-10) 1111111154411 1000005400 MIME-14'!)
14 I 00000005400 QW 0. 0000005400 0.0000005400 QW QWEM
24 I LWEM 1.0000005400 10000005400 1.5000545 1.5005545 10000005400
34 I 00000005400 0.0000005400 Q WEI-(X) 0.000005400 QW QW‘KXI
44 I OMB-(XI LW LWXIEKII 1000005400 1.0000005400 LWE-I-Q)
S4 I QWEHX) QWE-I-(D 00000005400 00000015400 QW QWEKKI
64 I QWEi-(X) 0.0000005400 00000005400 00000005400 LWM 0.5005545
74 I 00000005400 QW 00000005400 QQXXXXEMX) 1.0000005400 QWEM
S4 I OWEKXI QWEW 00000005400 00000005400 1.5500545 0.5005545
94 I QWEW QW 00000005400 0.000005400 LW 0.0000005400
104 I QWOOOEM QWEd-(X) 0.0555400 0.5005545 LWE-KXI MINNIE-10)
114 I 0.0000005400 0.0000005400 QWHX) QW 1.5555400 0.5050545
124 I 00000005400 00000005400 0.000005400 QW 10000005400 QW
134 I QW 0000005400 0000005400 QW LWM QWEAX)

3-2 99953352255552

39:35:25540

§§E§

 

 

 

 

Figure 36. (continued).

108

SIGNAL NAMES I
K007(25) Km K0070?) KW) K007(29) KOUI'OO)
QW OWN!) 0. W 0W 0.000005400 Q W
LWOEM 10000005400 1 .W 1W 1.5055400 1.5555400
00000005400 QW 00000005400 0000005400 QW 0 .(XXIOWE-HXI
LWEHXI LW 10000005400 10000015400 10000005400 1 .0000005400
00000005400 00000005400 0.0000005400 0.000005400 QW QWEKXI
10000005400 LW 1000005400 1000005400 LW LWEW
00000005400 00000015400 0000005400 0000005400 0.0000005400 0.0000005400
LWEW LW 10000005400 0000005400 1.0000005400 10000005400
00000005400 QW 0W QWM QW 00111111544!)
QWEM LWM 0.000005400 0000005400 QW 0.0000005400
QWEKD LW 00000005400 QWEd-(l) QW QWW
00000005400 1.0000005400 QWM QW 00000005400 QWEHI)
QW 10000005400 0.000005400 QW 00000015400 QW
0.0000005400 10000005400 0000005400 00000005400 00000005400 0.0000005400
QW 1000005400 0W QW 0W 0.000005400
I SIGNAL NAMES I
K00'I‘(31) K00'I'(32) W) K007(34) K007 (36)
QW 00000005400 0. W 0W 0.000005400 0W
LWOEAX) 1.0000005400 1 .W 1.0000005400 1000005400 LMM
QWEM QW 0. 000005400 0.000005400 QW QWE‘KXI
LWE‘KX) LW LW 1.000005400 10000005400 1.0000005400
0.0000005400 QW 0.000005400 0.000005400 QW 00000005400
LWEKD LW 10000005400 1.000005400 LWHX) 1.5500545
0.00me400 QW 0.0000005400 00000005400 QW QWHD
1.0000005400 LWM 10000005400 LWAXI LW LWEM
0. 0000005400 QW 00000005400 QW QW 0.0000005400
1 .0000005400 QW 1000005400 _ LW‘HII LW 1.0000005400
00000005400 QW 0. 0000005400 0.000005400 QW 00000005400
01111100514!) 00000015400 Q 000005400 0. WEI-(D 0.5005545 1.5005545
QWEHD 00000005400 Q 000005400 0. W44!) QW-I-(X) 1 .0000005400
0.0000005400 00000005400 0000005400 QWEd-(X) 0.0000005400 1.0055545
0.0500545 0W 0W QW 0W 1 .000005400

1-

SIGNALNAMEE I

(1118') I K007(37) K00'I‘(3S) K007(39) M40) KM“) K007 (42)

0 I W40) 0.000005400 QW 0W 0000005400 QW
4 I LW 10000005400 LW 1W 1.0005545 1.5005545
14 I 00000005400 QW 00000005400 0.000005400 QW 0.5500545
24 I 10000005400 LW 10000005400 LW LW 10000005400
34 I 0.0000005400 QW 0W 0000005400 0.0000005400 00000005400
44 I 10000005400 LW 1000005400 1.000005400 LW 10000005400
S4 I 0W QW 00000005400 0000005400 QW 00000005400
64 I 10000005400 LW 1111110844!) 10000015400 LW 1.0000005400
74 I 0.0000005400 QW 00000005400 0.000005400 QW 0.0000005400
S4 I 10000005400 LW 1000005400 1.000005400 101100544X) 1.0000005400
94 I 0.0505545 0.5000545 0000005400 0000005400 0.0000005400 00000005400
104 LW 10000005400 0000005400 LW 0.0055545 1.5000545

. QW QWM QWEHX)
0.0000005400 00000005400 00000015400 1.0000005400 00000005400 QW

00000005400 0.0000015400 WK!) LW 0W QW

I
1

E

1?

SIGNAL NAMEE I

I K0!I!‘(43) K007(44) W45) K007(46) K007(47) K007(4S)

00000005400 QW QW 00000005400
1.0000005400 10000005400 LW 1000005400
0 0000005400 QW 0W 0.000005400 0 W

:2-3

0000005400 QW

in
§

I . . 0000005400
24 I 1.0000005400 LW 1.000005400 LWW LW 1110108414!)
34 I 0.0000005400 QW 00000005400 QWM QWHD 0.5005545
44 I LIXIXXXIEHX) 1000005400 101110544!) LWM 1.0505545 1.0550545
S4 I 00000005400 QW 0.0000005400 00000005400 QW QWEM
64 I 1.0000005400 LW 1.0000005400 1.000005400 LW 1.0000005400
74 I 0.0000005400 QW 00000005400 00000005400 0.0000005400 0.0000005400
S4 I 1.0000005400 LW 10000005400 “BMW-14!) 1.0000005400 1.5005545
94 I 0.0000005400 0000005400 QW 0.000005400 QW 0.0000005400
104 I 1.0000005400 LWE'KXI 1.000005400 1.0000005400 MIME-14!) ”HINGE-14X)
114 I QWE'KX) 0W 0.000005400 QW 0 0000005400 0.5005545
:32: : 1.0505545 0.0050545 0W 00000005400 00000005400 0.0000005400

10000005400 0000005400 0W 00000005400

Figure 36. (continued).

109

 

713:5 I SIGNAL NAMES 4.
(NSI) I K0070) K0070) K0070) K007(4) K0076) K007(6)

0.0000005400 QM'FN 0.0005545 0.5005545 0.5005545
I LWHXI 1.0000005400 LW 10000005400 LW LWEM
LWOEM QWE+00 QW QW 00000005400 QW
10000005400 QW 0.0000005400 QW Q W44!) 0000005400
101100544!) QWXXEIN 0.0000005400 QW 0W QQXXIXIEI-(D
l.QXXXXIEd-(XI QQXXXXE-I-Q) QW QW 00000005400 QW
1010005114!) 0.000005400 QW 0.0000005400 0.000005400 0.0000005400
LONGER!) 0.000005400 QW QW 00000005400 01000111544!)
LWOEM QWHX) 0.0000005400 QW QW QWEHD
LWOE-HXI 0000005400 QW QW 000011005400 QW
1.5005545 0.5000545 QW QW QWEHX) 0.5000545
1.00me400 QW QQXXXXEKX) 0.0055545 0.0005545 0.000(XIOE+(XI
10000005400 QW QWW 0.0055545 0.5000545 QW
1.0055545 0.0055545 QWXEM QW 0.000005400 QW
LWEM QW 0.000005400 QW 0.5000545 0.0550545
1.5005545 00000005400 QWM QWE-HXI 0.000005400 QW
1.0000005400 QW 0.000005400 QW 0.000005400 00000005400
1.0000005400 QW QW QWE‘KX) 00000005400 0W
I IWEM QW 0.0555400 0.0055545 0000005400 00000005400
184 I 1011000544!) QWE‘KX) QWKXI QWM 0. 000005400 QW
194 I LWEAX) QWEKX) 0.000005400 OWN!) 0 .000005400 QW
N4 I 1.“!!me QW Q 000005400 0.0000005400 0. W 0W
214 I 10000005400 QW Q 000005400 00000005400 0W QW

I
5

§§§§§§22555522

‘—
23%

 

713:3 I SIGNAL NAMES I
(NS) I K007 (7) K007(S) K0079) K007 (10) K00701) K00702)

WEI-(X) 00000005400 QW 0W 0000005400 QW
1.5005545 1.0500545 1.5055400 1W 1.000005400 1 .000005400
QWEM 0. W 0000005400 0000005400 QW 0. 0000005400
2000000501 0 .0000005400 7.0000501 0.5005545 2W0! 0. WEI-(X)
0.0000005400 0. MBA!) QIXXXIXIEIQ) 0.000005400 QW 0. WEI-(X)
2500000501 0.0000005400 7W0! 0.0000005400 2000000501 QWEHX)
0.0000005400 QW 0000005400 0000005400 QW QWEKX)
2W0! QWEKX) 750000501 0.0000005400 2M0! 00000005400
00000005400 QW 0.0000005400 0000005400 QW 0.000005400
WE01 0. W40) 7W0! 00000005400 2500000501 0.0000005400
QWEHX) 0. MBA!) Q 0000005400 QWAX) QWM 0.0000005400
2.5(XIOOOE01 Q 000005400 7. SW01 QW W01 QW
QCXXXIOOE-I-(XI 0 .0000005400 0 .000005400 00000005400 0.0000005400 QW
2000000501 0 0000005400 7000000501 0011me 2500000501 00000005400
0.555545 0 0000005400 QWAX) QWEM 0.0000005400 QW
lSQDOOE01 0. WINE-100 7000000501 QW 2W0! QW
00110005414!) 0 000110544!) 0.000005400 QW 00000005400 QW
15411100801 0.5000545 7.0000501 QW W01 QWEKX)
0.0550545 0.5050545 0.5505400 QW 00000005400 QW
W01 0000005400 1W! 0.0000005400 W01 0.000005400
01110005114!) 00000005400 QWM QW 00000005400 QW
WE01 0. 000005400 7M0! QW 2W0! 00000005400
0.0000005400 0. 0000005400 QWEM 00000005400 00000005400 QW

§§i§§§§§§§§f§§ffff§ffig-

N
H
«b

Figure 37: Results of Three Stage Interconnection Pattern and Ramp Neuron.

110

r0145:

(NSH

§E§fritrtrfiztg_

§§§§§§§§

:3
;

=:2-3.§

§§§§§§§5eaaxzrx

-
Si

194I

§

214 I

 

 

SIGNAL NAMES
K00703) K00704) K0070S) K00706)
QWEHD 0. 0000005400 0. W 00000005400
10300054101 10000005400 1 .W 1.0000005400
0.0000005400 0. 0000005400 0. 0000005400 0.000005400
75WE01 75WE01 2W0! LW
0.0000005400 QWEAX) 00000005400 0.000005400
7W0! 7W0! 1500000501 1.0000005400
QQXIOQIEHD QW 0.000005400 QWKX)
75ME01 7W0! W01 LW
00000005400 QW 0.000005400 0000005400
7500000501 7W0! W01 l.QXXXXIE-I-Q)
QWFAQI 00000005400 0.5000545 0.0050545
7500000501 7W0! W01 IWHX)
QW 0.0000005400 0000005400 00000005400
7W0! 7W0! W0! 1.0000005400
QWEHX) Q 0000005400 QWW 0.0000005400
7.SQIOOOE01 7W0! 2M0! 1.0000005400
0011000541) 00000005400 QWM 0.0000005400
75WE01 750000501 W01 10000005400
QQXDOOEIQ) 0. 0000005400 QM“ 0.0000005400
7541000501 7W0! W01 101110310)
000000544X) 0000005400 0.000005400 QW
75WE01 7500001501 2500000501 10000005400
QW 00000005400 0W QW
I SIGNAL NAMES
K00709) K00'I'C20) K007 (21) K007(22)
00000005400 00000005400 QW 00000005400
1.5555400 1.0050545 1.0000005400 1 W
0.0000005400 0.0000005400 00000005400 00000005400
1.(X100(X)E+IXI LW 1.000005400 1011105114!)
QOQXXXJEd-(X) 0.0000005400 0.0000005400 0.0050545
1.00me400 MIME-10) 101100514X) 1.000005400
QWEH'D QWEM QWi-Q) 0000005400
IWEd-OO LW 1.0000005400 1000005400
OWL-1400 QW 0.0000005400 0000005400
LWE+00 1.0000005400 LWM 1.0050545
0.5555400 0.0550545 QW‘IQI 0.000005400
10000005400 LWEKX) 10000005400 LWOEI-(XI
00000054!) QW'KX) QW-HX) 000100844X)
1.0000005400 10000005400 1.5000545 1.5000545
011000054” Q WEI-Q) 0.000005400 QW
”KNOB-(X) 1.5050545 1.5000545 1.5005545
0.0505545 Q 0000005400 0.000005400 QW
1.0000005400 10000005400 LWM kW
QWEI-(X) QWE+00 0.000005400 0011100544!)
l.QXXIOOEHX) 1.0000005400 LWHD LW
0.0000005400 QWEI-(XI 0.000005400 QW
1110000544!) 10000005400 1000005400 LW
001000544X) 0000005400 QW QW

 

K00707) _ , K007 (IS)

0000005400 QWXIOOEHXI
1.0000005400 101100084X)
0.0550545 0.5005545
2W0! ZSQXDOB01
0.0000005400 QWEM
W01 1500000501
0.5500545 0.0505545
W01 lSQXBOE01
QW QWEI-(XI
W01 2.5MXXIE01
QWEHXI QWi-(XI
2M0! 2W0!
QWM 0.0000005400
2W0! 2500000501
QWEM 0011005110)
W01 W01
0.0000005400 0.0000005400
250000501 2W0!
0.0000005400 QWHX)
2W0! 2W0!
0.0000005400 QWEAD
W01 W01
0000005400 QW

 

Figure 37. (continued).

111

W)

K007(7.4)

QWEKD QW

LWM
0.0000005400

0. 0000015400

10000005400
0.5500545
1.5500545
0.5005545
1.5005545
00000005400
l.QXJOQIE-I-(X)
QWEW
LWEW
0.0000005400
10000005400
0.5005545
1.5005545
QWEM
1 .0000005400
0.0000005400
1 .WOEHX)
0.0000005400
LWHXI
QW
1.0000005400
QW

 

 

 

 

TIME . SIGNAL NAMBE I
I
(NS) I K007(25) K00706) K007 (27) K00'I‘(28) K007(Z9) K007(30)
I
0 I 00000005400 0.0000005400 Q W Q W 0000005400 QW
4 I 1.0000005400 1.0000005400 1 .W 1011110510) 1000005400 ”“me401
14 I QWEM QW 0 .0000005400 Q 000005400 0011000614!) QOQIOQIE+00
24 I LWE-HXI LWOEAD 1.0000005400 LW‘HX) l.QIXXXIE-o-Q) LOWER!)
34 I 0.0000005400 QW QWM 0000005400 QW 0.5500545
44 I 10110101544!) 1.0000005400 10000005400 1.000005400 1.0000005400 1.0000005400
S4 I 00000005400 QW QWEI-Q) Q 000005400 QWE-I-(X) QWM
64 I LWE-I-N LWEI-QI LW 1000005400 LWOEKX) LWB-I-Q)
74 I 0.0000005400 QWEAXI QWEKXI Q 000005400 QWEHX) 0.QX)0Q)E+(XI
S4 I 1.0000005400 1 .0000005400 1.000005400 1.0050545 1.0550545 LOWE-14X)
94 I Q 0000005400 0. MOE-10) 00000005400 QWM 0.0550545 0.5005545
104 I 1000005114!) 10000005400 LWIN LW 1.0000005400 1.0000005400
114 I Q 0000005400 0. 0000005400 0.000005400 QW 0.0000005400 QW
124 I 100100544X) 1.0000005400 1000005400 10000005400 MINDS-14X) 1.0000005400
134 I 0 .(XXXIOOEHX) 0.0000005400 0W QW 0.0000005400 QW
144 I 11100005414!) LWBd-Q) 1000005400 1.0000005400 LIXXXXDEM 1.0000005400
154 I Q 0000005400 0.0000005400 00000005400 00000005400 0.0000005400 QW
164 I 1110000514!) LW 1.000005400 10000005400 1.0000005400 1.0000005400
174 I 0.0000005400 Q W44!) 00000005400 QW 0011111844!) QWd-Q)
184 I 10111005440 10000005400 10000015400 l.QXXJOOE-I-(XI 10000005400 1.0000005400
194 I 00000005400 Q WEI-(X) Q 000005400 QW 0.0000005400 QWENXI
204 I 101000540) 1.0000005400 1.000005400 LW 1.5005545 1.5050545
214 I QW 0W 0W QW 0W QW
11145 I SIGNAL NAMES I
0118') I K007(31) K007(32) K007(33) K00'I‘(34) K00'I‘(3S) K00706)
0 I QW 0.0000005400 Q W 0W 0000005400 QW
4 I LWEM 10000005400 1 .W 1W 1000005400 l.QXXXXIE-I-(XI
14 I QWEHX) QW 0. 0000005400 00000015400 QW QWEAX)
24 I 1WE+QI 1.0055545 1.5050545 1000005400 1.1!!!)me LWEM
34 I 0.0000005400 QW 0.0000005400 0.000005400 QW 0 .0000005400
44 I l.QXJOQIEi-(XI LIXXXXIOEHX) 1.0000005400 10000005400 LW 1 .(XXXXXIE'I-Q)
S4 I QWEM QW 00000005400 Q 000005400 0.0000005400 QWE-Kl)
64 I 1.0000005400 1111me 1.0005545 1.5000545 1.0550545 1.5005545
74 I OWE-(XI QWE‘I-Q) 0.0000005400 0. 000005400 0.0000005400 0.0000005400
S4 I 10000005400 MIME-(II 1.000005400 LWEHXI 10000005400 1.0055545
94 I QQXIOIXIE-I-Q) QW 0.0000005400 0.000005400 00000005400 QWEM
104 I 1000005400 LWEW LWNXI LW LW'KXI l.QXIOOOE-I-(D
114 I 001100054X) 0.0000005400 0000005400 0.5500545 0.5555400 QW
124 I 101000510) 1.0550545 1.5505400 1.5500545 1.5005545 101000540)
134 I QQXDOOEAXI 0.0000005400 QWEi-W Q W QQXXXDE+00 QWEW
144 I 1011000544» 10000005400 10000005400 10000005400 LWEIOO 1 .0000005400
154 I 0.0000005400 0 “HIDE-100 0.000005400 Q W QWEKX) 0. MOE-10)
164 I 1.0000005400 1 WEI-(XI 1000005400 LWE-o-(X) LIXXXIQIE-KXI 1 .(XXXXIOEHXI
174 I 0.0000005400 Q WEI-(XI 0000005400 QW 0.0000005400 0. 0000005400
184 I 1.5555400 1.5050545 1.5000545 LW 101110844!) 1 .0000005400
194 I QWEI-(XI 0.0000005400 QWEM QW 0.0000005400 Q WIDE-Q)
204 I LWOEHX) IWM 1.000005400 l.QXXXXIE-I-Q) LWE-HXI 1.0000005400
214 I 00000005400 00000005400 00000015400 QW 00000005400 Q mm

Figure 37. (continued).

112

TIMBL

 

I
018') I KOUTOT) KOUTGS) KOU'I‘G9)

_§-_--____-_

-flflflfl--

N
3

G

trfizefffiiftr'fii‘

ffiflik

3-;

§§§§§"”""”

gerraextr2:.o

tffi

-§--_--_--__

N
H
b

w-

0.000000st
LWOEHX)
QWEM
1&1!”me
o.ooooooa+oo
Loooooomoo
o.ooooooa+oo
LWEM
o.ooooooa+oo
1 .ooooooawo
o.oomooE+oo
1 .ooooooaoo
QWE-I-(X)
1300000900
09000001900
l.QXXJOOEm
o.oooooos+oo
LWOE‘KX)
o.ooooooa+oo
1 .(XXXJOOE-I-(X)
QWEM
1.000000%
0W

KOUT(43)

ammonia-00
1.00000054-00

0900000300
1900000900
0000000300
l.QXXXDE‘I-m
QWEi-Q)
MIME-hm
o.oooooos+oo
LWEW

09000005400

11!!me
QWEd-(X)
1.0moooa+oo
o.ooooooa+oo
1moooos+oo
QWEM
1.ooooooa+oo
QWE-o-w
1.oooooos+oo
o.oooooos+oo
LWOEHI)
o.oooooos+oo

onoooooe+oo
10000005400
QW
1 .W
o.ooooooa+oo

030000002400

KOU'I‘(44)

0W
1.ooooool.=.+oo

QW
10000001900
o.ooooooa+oo
Lamoooswo
03000001900
LWEHD
QW
LW
o.ooooooe+oo
1.oooooos+oo
0900000300
1.ooomor~:+oo
0.0000(an
1 .oooooomoo
0W
1 .W-I-Q)
omoooomoo
LWMX)
o.ooooooa+oo
woooooaoo
o.ooooooa+oo

QW
1 .oooooomoo
0000000900
19000001900
09000001300
LWE-o-m
o.ooooooe+oo
1.0000001-:+oo
o.ooooooa+oo
LWEM
00000001900
19000005400
omoooamo
LWW
omoooomoo
LW-I-m
0000000500
Loooooaawo
o.oooooaa+oo
LW-I-w
QWXEm
mooooamo
omooormoo

KOUTGS)

QW
Looooooawo

o.oooooo.=.+oo
10000001900
o.oooooox~:+oo
1.ooomoa+oo
QWM
10000001900
QWM
LIXXXIDE-IQ)
o.ooomoa+oo
1 .ooooooawo
o.ooooooE+oo
1 .ooooooaoo
0.0m4-00
1 .ooooomawo
o.ooooom~:+oo
1 .ooooooewo
o.ooooooa+oo
LWM
onoooouawo
1mm
0mm

SIGNAL NAMES
KOUTGO)

SIGNAL NAMES

onoooooswo
LWE'I-m
QWW
lmoooos+oo
09000005400
Looooooewo
o.oooooos+oo
Loooooo£+oo
o.ooooom~:+oo
Looooooewo
o.oooooua+oo
l.QXWEm
QW
Loooooomoo
QW
LWEHI)
QW
LWEAX)
o.ooooooE+oo
Loooooomoo
QWEHX)
momma
QW

KOUT (46)

o.ooooooe+oo
mooooomoo

o.oooooa-:+oo
1 .oooooomoo
ammo
LWW
00000001900
1 .ooooooewo
o.ooooooa+oo
LWM
0.000000%»
1.ooooooa+oo
o.oooooos+oo
”RIDGE-rm

Figure 37. (continued).

113

KOU'I‘(4 1)

KOU'I‘(47)

09000008400
Loooooaa+oo
QWEHD
LW
o.ooooooe+oo
LWE-I-m
o.ooooooa+oo
LWE‘I-m
o.oooooos+oo
1.ooooooa+oo
QW
1900000900
03000001900
Loooooomoo
o.oooooos+oo
1.oooooos+oo
QWEM
Looooooa+oo
o.oooooo£+oo
MIME-IQ)
o.ooooooB+oo
10000005400
00000005410

KOUT (42)

QW
LWEHX)
o.ooooms+oo
Looooooawo
0000000900
1.oooooos+oo
o.oooooos+oo
LWEI-Q)
o.oooooor-:+oo
LWEM
011130me
1.0000001=.+oo
QQXXJOOEHI)

KWWIS)
QWEHX)

LWE-o-m
0900000500
1.ooooooI-:+oo
0.(XX)0Q)E+(X)
1.oooomE+oo
QWEi-(l)
LWEAX)
00000001900
1 .oooooom-oo
QWEAX)
MINDS-fl!)
QWEHD
LW
omoooomoo
LWE'I-Q)
QWE-o-Q)
1 .ooooooaoo
o.ooooooe+oo
LW
o.oooooo£+oo
LWEHD
QW

SIGNAL m g
KOUTG) KOU'I'(4) mum) zooms)
0000000300 0000000300 0.000000300 0000000300

 

KOUTO) K0070)

MIME-0Q)

fiiktKfiie

QW

1000000300
QW
QW
0.000000300
QW
QWE-I-(X)

LW
QWEi-(X)
QWE‘I-(X)
0.000000300
0.000000300
0000000300
0.000000300
QWEI-(X)
0000000300

1000000300
0.000000300
0.(X)0000£+(D
0.000000300
0.000000300
QWE-I-Q)
0.000000300
QW
0.000000300

flflhuuu—enuu

”flfluunfluuu

0.(X)(XXJOE+(X)
0000000300
0000000300
o00000m~3+00
QWKX)
0.000000300
0.00000ua+00
0000000300
QWd-(X)
0000000300
000000300

QWEKX)
0.000000300

:22!

"H-555-..“
g. .
é

1000000300

MIME-(D
1000000300

ffflik

000000300

SIGNAL NAMES 3
W) K001- (10) MI I)

0W 0W
1000000300 1W
Q 000000300 0000000300
SSW-01 0000000300
1500000301 QW
8m-Ol QW
SWOI 0.000000300
SSW-01 0000000300
SW-m QW
Sim-01 QW
SSW-01 0000000300
SSW-01 Q 000000300
SSW-01 Q 000000300
SSW-01 0000000300
SSW-01 0000000300
SSW-01 0.000000300
SSW-01 0000000300
SSW-01 0000000300
SSW-01 0.000000300
SSW-01 0000000300
SSW-01 0000000300

-%

KOU'I'OZ)

000000amo 0000000300
1.0000001=.+00 1000000300
0.000000300 0000000300
1500000301 0.000000300

0.000000300 0.000000300
0.000000300 0.000000300
0000000300 000000300
0.000000300 0.000000300
0.000000300 0000000300
0000000300 0.000000300
0.000000300 0.000000300
0.000000300 0000000300
0.000000300 0000000300
0.000000300 0.000000300
0.000000300 0000000300
0.000000300 0.000000300
0.000000300 0.000000300
0.000000300 0.000000300
0000000300 0.000000300
0.000000300 000(000300
0.000000300 0000000300

-§

2
§

’777§f7"

trackersreiffiirtrfizeo

-"7§777"’

Figure 38: Results of Three Stage Interconnection Pattern and Sigmoid Neuron.

114

TIME!—

SIGNAL NAMES I
I
(NS) l

KOU'I'OS) KOU'I‘OG) KOUTO‘I) KOUTOS)

QW QWEKXI 0000000300 QWOEKX)

1000000300 1000000300 1000000300 1000000300
0.000000300 0.000000300 0.000000300 QWEHX)

1500000301 LWHD 1500000301 LW-Ol
QWW 0000000300 0.000000300 0.000000300

0000000300 1 .000000300 LW-Ol 1500000301
0.000000300 0.000(00300 0.000000300 QWEW
0000000300 SSW-01 QWE-KX) 0.000000300
0000000300 SSW-01 0000000300 0000000300
QWM SSW-01 0.000000300 QWE-I—(D
0000000300 SSW-01 QWd-(D 0.000000300
0000000300 0500000501 0.000000300
0000000300 SSW-01 QWEW
0.00000(I-:+00 SSW-01 0.000000300
0.00000aa+00 SSW-01 QW
0000000300 SSW-01 0.000000300
QWM SSW-01 QWEW
QWEM SSW-01 QWEAX)
QWM SSW-01 0000000300
000000tn+00 SSW-01 0.000000300

ROW (13)

QWEM
1000000300
QWEHXI
SWE-Ol
Q WEI-(X)
S .0030008-01
Q WEI-(X)
QWEi-(X)
QWEM
QWEM
0.000000300
0000000300
0.000000300
0000000300
0.000000300
0.000000300
QWEAX)
0.000000300
0.000000300
0.000000300
0000000300

XOUI'(14)
0.000000300

errrtrrr:::g_

xifiiiiififi"

~

0000mm

 

reezxtu§i§frrirtriztg

--fl~fl
-————-—_-—

0.000000300
1000000300

LWE‘HX)
QWHD

Q 000000300
SSW-01
0.000000300
1500000301

QWE‘KX)
QWEAX)
0000000300
QWEd-(X)
0. W44!)
Q 000000300
0000000300
QWEM
Q W054“)
0000000300

KOU'I'(21)

LW

Q 000000300
1000000300
QWM
LWE‘HX)
0.000000300

W1

0000000300

1500mm

QWEKD
0000000300
0.000000300
QWKX)
QWM
0000000300
000000th
0.000000300
0000000300
QW
Qm-rm

SIGNAL NAMES
KOU'I‘(22)

W1

1 .W
000000300
LWHX)
0.00000ua+00
1000001300

KW)

QWKD 0000000300 000000t300
1000000300

QW
LWEM
0.000000300
1 .000000300

QWi-(XI

KOU'I' (24)

QW
1000000300

QWEAX)
LWEM
0.000000300
1.000000300
0000000300

1000000300 SSW-01

0.000000300 Q 000000300 0. 000000300

1500000301

0000000300
LW-Ol
0.000000300
0.000000300
QW
QM“)
0000000300
0.000000300
QW
QW
QW

Figure 38. (continued).

115

LW

1W1

LW
SSW-01

1.000000300
LWEM
1000000300
1000000300
LWE'IQ)
LWEM
1000000300
1000000300

1500000301
0.000000300

QW
QWEHD
QWEM
0.000000300
0.000000300
0.000000300
QWEd-Q)
QWEKD
QW
QW

 

TIME I SIGNAL NAMES I
I

(NS) I KOUT(25) KOUT(26) KOUT (27) KOUT(28) KOUT (29) KOU'I' (30)
I

0000000300 0.000000300 QW 0000000300 0000000300 QW
1000000300 1000000300 1000000300 1000mm 1000000300 1000000300
0.000000300 0.000000300 0000000300 0000000300 0.000000300 QWEd-(X)
URINE-(X) 1000000300 1.000000300 1000000300 1000000300 1000000300
0000000300 QW QWd-Q) 0000000300 QW 0.(XXXXX)E+(X)
1000000300 1.000000300 1000000300 LWHI) 1000000300 1.1!!!)me
0.000000300 QW 0000000300 0000000300 QW 0.000000300
1000000300 LW LWM 1000000300 LW 1000000300
0.000000300 QW 0.000000300 QW QW 0.000000300
85WE-01 HUMAN) 8.5m-Ol 5000000301 SSW-01 QWE-Ol
0.000000300 0000000300 0000000300 0000000300 QW QWEd-Q)
8W0! 1000000300 5000000501 1W1 8500000301 SWINE-01
0.000000300 0000000300 0000000300 QW QWM QW
15WE-01 350000301 0000000300 QW 1500000301 0.000000300
QW 1500000301 QW QW QW QW
0.000000300 8M0! QW 0000000300 0000000300 QW
0.000000300 8W0! 0.000000300 QWM 0.000000300 QWE-I-(X)
0.000000300 8m-Ol QW QWd-Q) 0.000000300 0.000000300
0.000000300 8.5m-Ol QW 0000000300 0000000300 QW
0.000000300 8m01 QW 0.000000300 0.000000300 0000000300
0.000000300 8.5m-Ol QW 0000000300 0000000300 0.000000300

iiii’ififiififtfifltkkho

'r-

SIGNALNAMES I

I KOUT(31) W32) KOUT (33) M34) KWI‘OS) KOUT(36)

0000000300 0000000300 QW 0W 0000000300 W
l.QXXXIOEHX) 1W LW 1W 1000mm 1.000000300
QWE'KX) QW 0000000300 QWW QW 0.000000300
1000000300 1000000300 LWM 1000000300 1000000300 1000000300
0.000000300 QW 0.000000300 QWM QW QWEd-(X)
1000000300 LW LW LW 1000000300 1000000300
QWEW 0.000000300 0.000000300 QW‘I-(X) QW 0000000300
1000000300 LW LWM 1.000000300 LW 1.000000300
0.000000300 0.000000300 QWM 0.000(00300 0.000000300 0.000000300
1000000300 LW 1W 1000000300 LW 1000000300
QWEHX) 0.000000300 0.000000300 0000000300 0.000000300 0000000300
1000000300 LWE‘KD 1000000300 1000000300 1000000300 l.QXXXXIEflX)
QOQDOOEM QWEM 0000000300 QWEHD QWEM QW
8.5WE-01 5000000301 8500000301 8W1 8.5WE-01 1000000300
0.000000300 0000000300 0.00000u300 QW 0.000000300 QW
8.5(XXIOOE-01 ISM-01 8500000301 ISM-01 50000001501 1000000300
QWOOOE-I-Q) 0000000300 0000000300 QW 0.0000(0300 QW
0.000000300 0.000(00300 QIXXXXXE-I-Q) QW 0000000300 1000000300
QWE‘HX) 0000000300 0.000000300 QW 0000000300 LW
0.000000300 0.000000300 0000001300 QWEHD 0000000300 1000000300
0000000300 0000000300 00000033400 QW 0000000300 LW

r::0utr§§§ff313338=:3_3

flu—oth-II‘
—-----—

Figure 38. (continued).

116

 

 

 

 

Figure 38. (continued).

117

TIME I SIGNAL NAMES I
I
(NS) I KOUT(37) KOU'I' (38) KOU'I'G9) KOUTGO) KOUT(41) KOUT (42)
I
0 I 0000000300 0000000300 W 0W 0W QW
4 I LWEHX) 1000000300 LW 1000000300 1000000300 1000000300
14 I 0000000300 QW 0000000300 Q 000000300 QW 0000000300
24 I 1000000300 LW LW 1000000300 1.000000300 1000000300
34 I 0.1!!!me QW 0111111124“) 0000000300 ‘ Q W 0.000000300
44 I IWB'HX) LW LW 1000000300 1 .W LWEW
S4 I OWEN!) QW QW 0W 00(0000300 0000000300
64 I 1000000300 1000000300 LW 1000000300 LW 1000000300
74 I OWEN!) QW QW 0000000300 QW 0.000000300
84 I 1000000300 LW 1000000300 1W LW LWEM
94 I OWEN!) QW 0000000300 0000000300 0000000300 QWE'I-(X)
104 I LW 1000000300 1000001300 LW 1000000300 1000000300
114 I QW 0W 000000c300 0.000000300 0000000300 0.000000300
124 I l.QDOOOE-I-Q) 1.000000300 1000001300 1000000300 1000000300 l.QXJOOOEHD
134 I 01113000344!) 0 .W 0.00000t300 QW QWEM 0. 000000300
144 I 1000000300 1000000300 1000000300 LW 1000000300 1100000300
154 I QOQJOOOEHD Q 000000300 0000000300 QWHXI 0. 000000300 0. 000000300
164 I 8.5WOOB-01 venous-01 SW1 1000000300 5000000301 8m-Ol
174 I 01111000240) QWd-N 0000000300 QW 0.000000300 0.000000300
184 I 0000000300 0000000300 0000000300 1000000300 0000000300 Q 000000300
194 I 0.000000300 0000000300 0.00000m+00 LW Q 000000300 0. W
TIME I SIGNAL NAMES I
I
(NS) I KOUT(43) KOUT(44) KOU'I‘(4S) KOUT(46) KOUT(47) KOUT (48)
0 I 0000000300 0.000000300 0.000000300 0000000300 0000000300 QW
4 l' l.QXIXJOE-I-Q) 1000000300 LW 1000000300 LW‘IN 1000000300
14 I QWEM Q W QWEd-Q) 0.00000(E+00 0.000000300 0000000300
24 I l.Q)OOIXJEW 1000000300 1.“!!!an 1.W00(E+(X) l.QHXIOE-I-(X) l.Q)OOQJBHX)
34 I 0.000000300 0.000000300 QWOEI-(X) QWM 0.000000300 0.000000300
44 I LWEW LW 1000000300 LIXXXXXE‘KXJ LW MINDS-HI)
S4 I 0.000000300 QW QWE‘KXI o00000m=.+00 0000000300 0.000000st
64 I “KNEW LW 1000000300 1000000300 LW l.QXXXDE-o-(X)
74 I 0.000000300 0000000300 QWEd-N 0000000300 QW QWEKX)
84 I 1000000300 LW 1000000300 1000000300 LW 1000000300
94 I 0.000000300 QW 0000000300 0.000000300 0.000000300 0 “WEI-(X)
104 I l.QXXJOOB‘I-(X) 1000000300 LWM 1.WOE+(X) 1000000300 1000000300
114 I 0.000000300 0.000000300 0000000300 QIXXDOOW 0 .000000300 Q WEI-(XI
124 l 1000000300 1000000300 1.000000300 1000000300 l.QXXXDE-KX) LWKX)
134 I 0.000000300 0000000300 0000000300 QW QWEd-Q) Q 000000300
144 I LWOEMX) 1W 1000000300 LWEW 1000000300 1 .WOE-I-(D
154 l. QWEM 0.000000300 0000000300 QW 0. WEI-(X) 0000000300
164 I l.QXXXJOE-I-(X) LWEW 1.00000aa400 LWEHX) LWEKX) 1000000300
174 l 0.000000300 0.000000300 0000000300 0.000000300 0000000300 QWEi-Q)
184 I 1000000300 0000000300 0.000000300 0000000300 0000000300 QW
194 I 1000000300 0000000300 000000tn+oo QW 0000000300 0.000000300

 

 

11MB! SIGNAL NAMES e
I
ms.) I room) room) now-(3) Koo-0(4) xou'rm Korma)

0 I 0000000300 0W QW 0.000000300 QW 0mm
4 I 100000(300 1W 1000000300 1000000300 MINER!) 1000000300
1000000300 QW QW QW 0.000000300 0.000000300
1000000300 0000000300 QW 0.000000300 0000000300 QWE-I-(X)
LWEW QW‘KX) QWEKD 0000000300 QWEKX) QWEM
LWOEW 0000000300 0.000000300 QW-I-(XI QWi-(XI 000000300
LWEM 0000000300 0.000000300 QM“) 0000000300 000000300
1000000300 000mm QW QW 0.000000300 QW
1000000300 0mm QW QW 0000000300 0.000000300
LWEM 000000300 QW 0000000300 0.000000300 QW
LW 0W QW QW QWW QM“)

“are“:

 

 

-Q-?
§
I
i
E

KOUTG) XOUI' (10) KOUTO 1) KOU'1‘(12)

0000000300 0mm QW 0W 0000(00300 QW
. . . LMM LW
OWE-KI) QW QWM 0000000300 0000000300 QWEAX)
0000000300 0000000300 1000000300 000000300 QW 0000000300
0000000300 QW 1000000300 0000000300 QW 0.000000300
0000000300 QW LWOE‘KX) 0000000300 QWEI-(l) 0.000000300
0000000300 QW 1000000300 0000000300 QW 0000000300
0000000300 QW 1W 0W QW 0000000300
0.000000300 QW 1000000300 0000001300 QW 0000000300
0.000000300 0.000000300 1.000000300 0000000300 QW 0.000000300
0000000300 0W 1000100300 0.00000a~:+00 QW 0000000300

E
E
E
I

fffliitffl'EWG

 

KOUTOS) M14) KOUTGS) KOUTO6) KOUTO7) KOU'I'(18)

0000000300 0000000300 QW 0W 0.00000t300 QW
1WE+M 100mm 1000000300 1W 1.00000tmoo 1.000000300
0000000300 QW 0000000300 0000000300 QW 0.000000300
1000000300 LW 1000000300 1000000300 MIME-H!) 1000000300
0000000300 QW 0.000000300 0.0000m300 QWd-(D 0.000000300
QWEM QWHX) 0.000000300 1000000300 0.000000300 0.000000300
QWEKXJ QW 0000000300 100000(£+00 0.000000300 0.000000300
0.000000300 QW 0000000300 1000000300 QW 0000000300
0000000300 QW 0000000300 1000000300 QW 0000000300
0000000300 QW 0000000300 1000mm 0000000300 0000000300
0000000300 QW 0W 1000001300 QW 000000300

_§_g
I
E

frflflttfifii'fl’

Figure 39: Results of Four Stage Interconnection Pattern with Binary Neuron.

118

22:21:“:

2232333303396

1

2-3

K0011”)

KOUTQI)

Q 000000300
1.11me
0000000300
1000000300
0000000300
0000000300
0000000300
0W
0000000300
QW

 

1-

0000000300

Q WEI“)
LWEW
QWEM
1000000300
0.000000300
0.000000300
0000000300
0000000300
0000000300

SIGNAL NAMBE

SIGNAL NAMEE
' KOUTO3)

KOUTQZ)
W
1W
Q W

1000000300
Q 000mm
0.00000(£+00
QW
0W

000000us+00
0W

M)

0m

0000000300
1 .mm
0000000300
100000t300
0.00000(£+00
QWW
0.00000t300
0000000300
0000001300

Figure 39. (continued).

119

KOUTGS)

KW)

000000aa+00
100000aa+00
QW
1.1!!me
0.000000300
LW
MIME-04X)
1000000300
1000000300
LW
LW

KOUTQA)

QW
LW
QWd-(X)
1000000300
0000000300
0000000300
0000000300
0.0000001mo
QWEM
0000000300
0W

0000000300 QW

1000000300

1.000000300
0000000300

Q WEI-(X)
Q 000000300
Q 000000300

QW

1.000000300
0.000000300
1000000300
0000000300
1000000300
QWE‘KX)
1000000300
1.000000300
1 .000000300
1000000300

 

 

 

 

TIME I SIGNAL NAMES I
I

(NS) I KOIJT(37) KOUI‘G8) KOU'I‘(39) W40) M41) KOUT (42)

0 I 0000000300 0000000300 QW 0W 0000mm QW
4 I LWOEKX) 1000000300 LW LW ”Hm-om 1000000300
14 I 0000000300 0.0(0000300 0.000000300 0000000300 QW 0.000000300
24 I 1000000300 1000000300 1000000300 1000000300 LW LWKX)
34 I OWE-(X) QW QWW QW QW QWW
44 I 1000000300 1.1!me 1000000300 1000000300 LW 1000000300
S4 I 0000000300 0.000000300 0000000300 0000000300 QW 0000000300
64 I 1000000300 LW 100mm 1000000300 LW 1000000300
74 I 0.000000300 0000000300 0000000240 QWEM QW 0.000000300
84 I 0.000000300 0.000000300 0W 1mm W 0000000300
94 .I 0W W 0000000300 1W W 0000000300
TIME I SIGNAL NAMEE I

I

(NS') I KOUT(43) W44) KOUT(4S) KOU'l‘(46) Koo-m7) XOUT(48)

0 I 0.00am300 0000000300 QW 0W 0000001300 QW
4 I. 1000000300 1000000300 LW 1000000300 1W+N 1000000300
14 I QWEHX) 0.000000300 0.000000300 0000000300 QW 0000003300
24 I 1000000300 1000000300 1000000300 1000000300 1000000300 1000000300
34 I QWEM QW 0.000000300 0.00000<300 0000000300 0000000300
44 I LWE'I-(X) 1000000300 1000000300 1000000300 1000000300 1000000300
S4 I QWE-I-(X) QW 0.000000300 0000000300 QW OWEN!)
64 I MIME-(X) 1000000300 1000000300 1000000300 MIME-HI) LWEM
74 I QWE-I-(X) 0.000000300 0.000000300 0000000300 0.000000300 0.000000300
84 I 1000000300 QW 0000000300 000000a=.+00 QW 0.000000300
94 I 1000000300 QW 0000000300 0.00000a=.+00 QW 0.000000300

Figure 39. (continued).

120

n?“
«9!
I

KOUTU)

I
é

§§§§§§§§§§§3333332§3>

N
H
‘

-2

-3

guide-o.—
----—--

3
;

D-I-lu-oo-l-e.‘

rrrarrrr

rererrrrrao

:9

5

g

1000000300
1 .OOQIOOEHX)
1. WORK!)
1.000000300
1000000300
l.QXXIOOEHX)
1000000300
1. WEI-(X)
1 .(XXDOOE-I-(XI
1000000300
LWE-l-(X)
1000000300
1 .WE-I-(X)
1 .(XJOQXIEM
1000000300

1 000000300
1000000300
LWEAX)
1WE+OO
1WE+N

'F'

QWEM
1000000300
0000000300
0000000300
0. WEI-(X)
0.000000300
QWEW
QWEHX)
QWEHX)
0.000000300
0 “HIDE-00
O. OWOOOEHX)
Q 000000300
0. WOOOE-I-OO
Q OWOOEHXI
0. WOOOEAX)
Q ”00050-00
0. WEI-(X)
0. 000000300
Q WEI-(X)
Q 000000300
Q WEI-(X)
Q WORM

mmmn

Q 000000300
1000000300
0. 000000300
0000000300
QWW
0.000000300
0.00000(B+00
QWW
0000000300
QWM
0000000300
0.000000300
QWEM
O.(XXXXXIE+(X)
0.000000300
0.000000300
0.000000300
0.000000300
0.000000300
0.000000300
QW
QWEHX)
QW

0. 000000300
Q WEI-(X)
Q 000000300

KOUTG)

smmuunms
awn»

0000000300 0.000000300

LWEAD

QWEHXI
QIXXDOOEM
0.000000300
QWE‘I-Q)
QWEHX)
0.000000300
QOQIIOOE‘I-(X)
0000000300
QW
0.000000300
0.000000300
0.00000IE+00
0000000300
QWW
0000000300
000000300
QIXXXXXEM
QWM
QIXIOIXXE'I-(XI
0000000300
0000000300

ROI-m9)

SIGNAL NAMES

1000000300

0000000300
0.000000300
0.000000300
QWE-I-(X)
QWEMXI
QWE-I-(XI
0.000000300
0.000000300
0000000300

QWEHX)
QWEW
0.000000300
QWEM
0.000000300
QWEHX)
QWEM
QWE-HX)
0.000000300
0.000000300
0.000000300
0000000300

nmum)

“WK”

QW
1000000300
0 .(XXXXJOE‘KX)
Q 000000300
0.000000300
QOIXXDOEI-(X)
0. 000000300
0. WEI-(X)
Q WEI-(X)
0.000000300
0.000000300
QW-FQ)
0000000300
QWKX)
QWM
QWHXI
0000000300
0000000300
QQXXXJOEM
000000300
0.000000300
0. WEI-(X)
Q 000000300

nmuu)

HNNQ

0.000000300
LWEKXI
0. 000000300
0. WEAK)
O. WOOOE-I-(D
0.000000300
0113000084!)
QQJOOIXIE-I-(X)
QQXXJOOEHXI
0. WEI-Q)
Q (”GONE-IQ)
0. WEAK)
O .QXXXXJEHX)
QWEW
0. 000000300
0 .WOEHXI
0. WEI-(X)
O. WEI-(X)
Q 000000300
Q MOWER!)

0 000000300
Q 000000300

ammu)

Q 000000300
1.”!me
Q 000000300
ISM-01
7500000301
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
ISM-01
mom-01
ISM-01
ISM-01
ISM-01

ml

0.000000300
1W
0000000300
QQXXDOEI-(X)
QWEHX)
QWEHX)
0.000000300
QW
0.000000300
QWEM
QQXXJOOEA-(X)
QWE-I-(X)
0000000300
QWEHX)
0.000000300
0.000000300
0.000000300
0.000000300
0.000000300
QWE'KX)
0.000000300
Q 000000300
Q 0000mm

0.00000a-:+00
1000000300
0. 000000300
0 .000000300
0 .WEM
QWE-I-(XI
0.000000300
QIXXXXDE-I-(XI
QWEHX)
0000000300
0.000000300
0000000300
0.000000300
QWEHD
QIXXXIOOE+QI
QW
QIXXXJOOEHI)
QWEHX)
0.000000300
0.WE+(D
MIME-Kl)
QWEM
QW

0. MEN!)
1 000000.300
O. WEM
O.(XX)000E+(X)
0.(XI0000E+Q)
QQIOOOOE+QI
QQIOOOOE-HI)
0.000000300
QQXIOOOEHXI
0.000000300
QWOOOEHX)
O.(XXX)00E+(XI
0.000000300
QWEI-(X)
0.000000300
QOWOOEHX)
QQXDOOE'I-(XI
0.000000300
QWEHX)
0000000300
0 .WOEHXI
Q 000000300
0000000300

Figure 40: Results of Four Stage Interconnection Pattern with Ramp Neuron.

 

 

 

 

Figure 40. (continued).

122

TIME L SIGNAL NAMES E
KOUTOB) KOU'1'(14) KOUTOS) KOUT(16) KOUTO‘I) KOUTOS)
0 II aooomom-oo ooooooomoo Q W OW ammo QW
4 I l.QXXXXJEm LWE-Im LW LW Imoooomoo LW
14 I o.ooooooa+oo o.ooooooa+oo Q oooooomoo o.ooooom+oo QW omoooomoo
34 I 1.ooooooE+oo LW 75000001501 LW 7500000301 7500000501
34 I o.oooooox-:+oo QWE-o-Q) QM“ 00000005400 0. W 0. WEI-(X)
44 I W01 5.WE01 oooooooa-oo 1W 1W1 omooooawo
54 I o.oooooos+oo QW o.oooooos+oo ammo Q W o.ooooooa+oo
64 I W01 momma-01 QW Looooooa+oo W01 o.ooooool-:+oo
74 I o.ooooooa+oo o.oooooo'='+oo QWEM QWM QW 0900000300
84 I W01 QWE01 QWHD LWEM 2W1 o.ooooooa+oo
94 I QWEAI) 09000001900 o.ooooooe+oo 09000009300 o.ooooooe+oo o.ooooooE+oo
104 I 1511000501 SW01 o.ooooooa+oo 1.000000% W01 QW
114 I o.ooooooa+oo Q oooooo£+oo cocoon-2400 QW 000000015400 o.ooooooa+oo
121 I 1501100501 5 mom-01 omooooewo Loooooos+oo W01 QWEd-(X)
134 I QWEW omoooos+oo omoooomoo o.oooooos+oo o.oooom£+oo QWEM
144 I 2.5(XJOOOE01 SW01 09000005410 LW 2W1 QW
154 I QWOOOEd-Q) 0. 0000001900 o.ooooooa+oo QW o.oooooos+oo QW
164 I 2.5(XIOOOE01 5 .ooooooe-m o.ooooooE+oo l.QXXXXIE'o-m W01 o.ooooooa+oo
174 I QWE-I-(X) Q oooooomoo o.oooooos+oo QWHD o.ooooooE+oo o.oooooor~:+oo
184 I 1500000501 1000000501 QWEM LWM W01 0000000900
194 I QWEHX) Q WEI-Q) 09000005400 QW QW QW
204 I 2.5(XIOOOE01 5.1XXXXIIE01 o.ooooooB+oo momma-co W01 0000000900
214 I QW 0W 0W ammo o.ooooooa+oo QW
“11313 .L SIGNAL NAMES 5
(NS) I KOU'I'(19) KOUI‘QO) KOUT (21) KOUT(22) [(00103) KOUTQA)
I
o I- 09000005400 omoooomoo Q W 0W 0mm QW
4 I Loooooos+oo mooooos+oo Loooooos+oo 1W moooouswo 1.00000054-00
14 I Q oooooos+oo Qme 0 .oooooomoo o.ooooooa+oo QWHX) onoooooE+oo
24 I LWE‘I-m 1.0000001=.+oo 15000001501 7.5(XXXIOE01 Loooooosm LWE-I-(l)
34 I o.oooomE+oo QWEM o.ooooooa+oo o.oooooa£+oo o.ooooooa+oo 0.0000005“)
44 I W01 smooooa-m QWEHD W01 10000005400 momma-01
54 I Q oooooos+oo o.oooooos+oo QWd-(X) omoooaa+oo common-+00 o.ooooooI-:+oo
64 I ZWE01 10000001501 o.oooooos+oo W01 19000003900 SW01
74 I Q (BOOGIE-(X) o.ooooooa+oo 0 .ooooooswo o.ooooooa+oo QW QWEHXI
84 I W01 QWE01 QW WE01 1900000300 5.(XW)E01
94 I o.oooooor-:+oo QWE-I-Q) Q WEI-(X) o.ooooooa+oo QWEHX) QWEKX)
104 I 2500000301 5.m01 0 .oooooomoo 2.5QDME01 LWBW 1000000301
114 I QWOOOEAX) o.oooooos+oo o.oooooos+oo QW o.oooooos+oo QW
124 I 2W0] 1000000501 Qm-KX) 2500000301 Loooooomoo 5.WE01
134 I Q WOOOEHXI Qmm o.oooooo:=.+oo QW o.ooooooa+oo o.oooooos+oo
144 I 2. W01 5.00000on 0.0000005+oo 2W1 moooooawo 1000000501
154 I Q oooooom-oo QWEKX) omooooem QW 0.0000(an 00000003400
164 I 1500000301 1000000501 QWKX) 2.5MXXIE01 Looooooewo 1000000501
174 I o. oooooo£+oo o.ooomo15+oo o.oooooa3+oo QWE-I-Q) 00000005400 QW
184 I lSWE01 momma-01 o.oooooo£+oo W01 mooooaa+oo 5.WE01
194 I Q ooooooewo 00000001900 omoooaswo Qoooooomoo ooooooo£+oo QW
204 I WE01 SW01 0mm W01 Loooooaa+oo 1W]
214 I QW ammo omoooumo QW 0mm 00000001900

 

 

Figure 40. (continued).

123

ms.) I KOU'ITZS) KOUTQG) KOUTC") K0011”) [(00109) KOUTGO)
O I aooooooswo ammo 03000005410 announce-00 omoooaawo aooooooawo
4 I momma-poo LWEHX) 1.“!!me LWEW 1.ooooooe+oo LWEW
14 I o.oooooos+oo o.oooooo:-:+oo o.ooooooa+oo o.ooooooe+oo o.oooooos+oo QWEM
24 I LOWE-IQ) 1.0000001=.+oo Loooooos+oo LWM l.QXmOE-HXJ 1.oooooox-:+oo
34 I o.ooooooa+oo QW o.oooooon+oo o.ooooooa+oo QW Q WEI-(X)
44 I 1000001841) 1900000900 LW Looooomwo LWHX) LWEM
54 I Q 0000001900 uncommon o.ooooooe+oo o.ooooooa+oo QWW Q WK!)
64 I 10000001900 LW LWEHX) LW-I-m 11111110544!) 1 .ooooooaoo
74 I Qmem o.oooooos+oo Q WIDE-IQ) o.oooooa-:+oo QWEKX) 0.0100109“)
84 I Loooooomoo LWEKD Looooooa+oo wooooomoo LWEd-m l.QXXXXJE-I-(X)
94 I o.ooooooa+oo QWEHD o.oooooou-:+oo ammo QWEHX) common-+00
104 I l.QDOOOE-I-(X) LWEW LWM Loooooos+oo LWEKX) l.QXXXXIEAXI
114 I QWEHD QWi-(X) o.oooowe+oo oooooooE-Ioo o.ooooooa+oo QWE-I-Q)
124 I l.QXIOOOE-I-m LWE'I-(XI LWM Loooooomoo wooooomoo Loooooos+oo
134 I QOQXIOOE-I-w o.ooooooE+oo omooowwo QW o.ooooooa+oo 0300000900
144 I 1.0000001:-:+oo LWE‘HX) LWEM LWNX) LW-om LWOEM
154 I 03000001900 Q 000000300 QWi-(D Q ooooooa+oo QWE‘HD 01!!!me
164 I Loooooomoo 19000001900 mooooomoo 1 .oooooom-oo Looooooewo LWEHD
174 00000005400 QWEKX) o.ooooom+oo uooooooa+oo o.ooooooe+oo QW
184 I l.QXXJOOEKX) l.QIXXIOE-rm Looooooswo 1.000000F.+oo LWDEM 1300000900
194 I QIXXXJOOEW QWEM o.ooooooa+oo o.ooooooa+oo o.ooooom-:+oo 0. WEI-(XI
204 I LUMBER!) LWE-bm LWM LWIEM moooooa+oo LWEM
214 I o.ooooorm+oo Momma OW W 000mm QW
TnfE .L SIGNAL NAMES I
(NS) I [091131) KOUTGZ) KOU'!‘(33) KOUTCM) KOUTGS) KOUTGG)
O 'I ooooooomoo 0000000300 Q oooooomoo Q 0000001900 0mm uoooooom-oo
4 I 1 oooooom-oo 1111111154!) 1 .0000me LW Locum-mo MIME-rm
14 I Q WEI-(X) QW Q oooooomoo o.ooooool-:+oo QW 0300000900
24 I LWOQIEM l.QXIOOOE-I-Q) Loooooomoo momma Lamooom l.QXXXXIE‘KXI
34 I 0 .oooooomoo QW o.ooooooa+oo QWW 0.00000054-00 omoooomoo
44 I LWEi-w LWE-o-Q) 1.oooooor~:+oo mooooomoo LWE-o-(l) l.QXXXDE'o-Q)
54 I o.oooooos+oo QW 0900000900 o.ooooous+oo QW QWEd-(X)
64 I l.QXIOQJE-I-m ”MIME-rm 1 WEI-(X) LW'KXI LWE-o-m 1 .(XDOIDEM
74 I 00000001900 QWE-I-Q) Q WEI-Q) 0900000900 QW Q WEN!)
84 I LIXXXXXIEW LWEHX) LWEM 1.ooooom+oo LWEM 1900000900
94 I o.ooooooa+oo 0.000000134-00 Q oooooos+oo QWM o.ooooooa+oo Q WEI-(X)
104 I LWOOOEHX) LMOE-I-m lmooooawo l.QXXXDE'I-m Looooooe+oo Loooooomoo
. 114 I QWOEHX) o.oooooos+oo 0000000900 QWEW QWE-KX) o.ooooooa+oo
124 I l.QXXIOOE-KX) Looooooa+oo 1000mm momma-+00 l.QXXXDE-Im 1.“!!me
134 I QWXIOOEAX) o.oooooo;-:+oo o.ooooooa+oo QW QWEM QWKD
144 I LIXXDOOE-I-m 1.000000E+oo 19000001900 1.0000001-:+oo LWM LWEW
154 I QWE-I-(X) QWEHX) o.ooooooa+oo QWEHX) o.ooooooE+oo QW
164 I l.QXDOOE-I-w Looooooawo l.de-m MIME-04X) MIME-hm LW
174 I QWXIOOE'I-m QWE‘I-w o.oooooa-:+oo QW o.oooooos+oo 03000005400
184 I l.QXIOOOE-I-m 1000000300 LW-KX) LWEHD LW‘KXI 1.1!!wa
194 I o.oooooos+oo QW‘IN ommoaawo QW o.oooooo£+oo omooooawo
204 I LWEHD 10000001900 10000013400 LWE'I-m 1W4“) LW
214 I QW o.ooooooa+oo ammo ammo 0.0000005+oo QW

 

 

 

 

Figure 40. (continued).

124

TIME L SIGNAL NAMES I
I
018') I K00'1‘I37) K00'1'(38) K00'I'(39) [001140) K0011“) KO0T(42)
O I QW omooommo Q «momma 0W omoooua+oo o.ooooooa+oo
4 I Loooooomoo 1000000900 1 .W LW mooooamoo l.QXXINE-I-Q)
14 I QWEM QW Q oooooomoo omooooawo 0.00000054-00 QWE-I-(X)
24 I Loooooos+oo LW LW Loooooaawo 1111100544!) 1.ooooooe+oo
34 I QW QW OW QW o.ooooooe-oo omoooomoo
44 I Loooowmoo LW mooooomoo wooooaswo LWEM 1.0000005+oo
54 I QWHX) QW o.ooomoe+oo o.oooooo£+oo QW o.oooooos+oo
64 I l.QXXXDE-I-(XI LWE-om mooooomoo 1900000900 10000001900 Looooooewo
74 I o.ooooooE+oo QW 0W omoooaawo o.ooooooE+oo o.ooooooa+oo
84 I LW‘I‘N LW LW Looooamoo 1W Loooooomoo
94 I QWEd-m QWi-N omoooomoo ammo o.a100001=.+oo QWEM
104 I hm LW 1.ooooo(£+oo LW LWEM 1000000900
114 I QW QW Q W o.oooooos+oo omoooomoo QWOOOEHX)
124 I 1010000544!) moooooe+oo LWM Loooooo£+oo 1000000300 ”HINGE-(X)
134 I QW 0W omoootmoo QW o.ooooooa+oo o.ooooooa+oo
144 I Lamoooam 1.0000005+00 woooomwo LWOE-I-(X) LWEKX) l.QXXXXIE-o-w
154 I QW 0900000300 0mm QW o.oooooos+oo QW
164 I LW LW 1W LW LooooooE+oo l.QJOIXDE'I-QI
174 I QWE-I-(D omooommo onooootmoo QW o.oooooo£+oo 0.(XXXXX)E+(D
184 I Looooooa+oo 10000002400 19000005400 LW 19000005410 l.QXJOOOE-rm
194 I o.ooooooa+oo QW ammo QW ammo o.ooooooa+oo
204 I LW 1W 19000005400 LW Imooomam LW
214 I QW OW OW amen o.oooooon+oo QW
TIME .L WA]. NAMES I
I
(148') I [(001143) KM“) KGJTGS) [001(46) W47) K00'1'(48)
O I aoooooomoo omoooomoo Q W W OW QW
4 I- Loooooos+oo LW-Im LW LW moooooewo 1.000000%
14 I 000000013410 QW Q 0000005400 OW QW-O-(XI 00000001900
24 l 1.oooooo£+oo LW looooom-zwo 1000001300 LW LWEM
34 I 0.00000054-00 QW o.ooooooa+oo o.oooooaa+oo QW o.ooooooewo
44 I l.QXIOIXIE-I-(X) 1.ooooool-:+oo 1.000000%» l.QXXXXE-I-(X) LW LWKX)
54 I QWEM QW QW o.oooooa=.+oo QW 0000000900
64 I Loooooos+oo hm 1900000900 LW'I-m LWE-I-Q) LWEAX)
74 I o.oooooos+oo QW 09000005410 QWM QW 0.0000005+oo
84 I 19000005400 LW 19000008400 Loooooaa+oo LW 1900000300
94 I o.oooomE+oo QW omoooomoo o.oooooos+oo QW 0900000900
104 I l.QIXXIOE-I-(X) Looooooaoo mooommoo Looooooewo LWW LWOEHD
114 I QW omooommo Q oooooas+oo o.ooooooa+oo Q oooooomoo QW
124 I LW 13000001900 1 oooooamo LW 19000001900 LW
134 I Q MOE-H!) o.oooooos+oo Q oooooomoo QW ammo o.oooooos+oo
144 I 1 .ooooooawo LIXXXXXIE-rm Looooouawo Looooooa+oo LWE'I-m 1 .(XXXXXIE‘I-Q)
154 I Q MOE-HI) 00000001900 o.oooooaa+oo 0. W QWE-I-(X) Q 000000900
164 I 10000001900 Loooooomoo Lamond-300 1.oooooo£+oo 1.000000st 1 .oooooom-oo
174 I Q MOE-«(D o.ooooooa+oo o.ooooom+oo Q 0000001900 QWEKX) QWEND
184 I 1 .oooooom-oo LIXXXmE-I-Q) woooouawo hm LWEM LW
194 I Q WEI-(XI OWEN!) Q oooooua+oo QW 09000005400 o.ooooooe+oo
204 I 1.WOE+M ”HIKE-IN wooooaawo 1000000300 mooooomoo 1.oooomE+oo
214 I Q oooooomoo omooooaoo ammo QW 00000001900 QW

11MB! SIGNALNAMES I

K0010) K001‘(4) K001‘(5)

OI o.oooooas+oo 0W QW ooooooomoo

. o.ooooooa+oo o.ooooooa+oo
Looooooa+oo LW LW Looooooswo LW LWE'HX)
l.QHIOOE-tm omooooewo QW-HX) QW o.oooooo:=.+oo QWEW

LIXXXIOOEM o.ooooom+oo QW QW QWEd-(X) QWE‘o-Q)
LWEHX) o.ooooocmoo QW omoooosm QWEM 01000me
LWE-I-m QW'KX) QW QW 090000013400 QQXXIOOEAD
l.QXXIOOE-I-(X) o.oooooaa+oo QW QW 0000000900 0110000544!)
1.(XXX)OOE+(X) o.ooooooa+oo QW QW QWE-I-(X) o.ooooooE+oo
LWOE'I-(X) o.ooooooe+oo QW QW o.ooooooa+oo QWOOIXIEi-(X)
l.QXIJOOEm QW‘I-(XI QW QWEM o.oooooos+oo o.oooooos+oo
Looooooewo o.ooooooa+oo QW QW QWEM 0900000300
1900000500 0.0mm o.ooooooa+oo QW o.ooooooa+oo 00000003400
1.(XIOOOOE+OO o.oooooos+oo 0000101544!) 0100100540) o.ooooooa+oo 0000000300
1WE+QI o.ooooooa+oo o.oooooua+oo QW 000000015400 000000015400
19000001900 QW omooamoo QW o.ooooooa+oo 0W

I
(NS') I K0010) K0010) K0016)

§§§§rrrcrrrrzt

:3

SIGNAL NAMES I
K001 (9)

s
E

K00'1‘00) K00101) K00102)

aooooooE+oo

--

Loooooos+oo
omoooos+oo
QWEi-(X)
o.ooooool.=.+oo
OWE-om
QWE'I-(X)
QWEM
QWEW
QIXXXXXIE-I-(D
QWEi-(X)
o.oooooos+oo
o.oooooos+oo

§§§§rrrextrexso

QW
LW
o.ooooooa+oo
“cocoon-01
85010501
8m01
8m01
8W0]
8W0]
8m0l
8m01
8m01
8m01
8m01
W01

onoooooawo
1.ooooooe+oo
o.ooooooa+oo
QWEM
omooooawo
o.ooooooE+oo
o.oooooon=.+oo
QWEd-(X)
o.oooooo£+oo
o.oooooos+oo
QWE‘I-m
QW
QW
o.oooooos+oo
0W

QW
l.QXXIOOE-I-(l)
o.ooooooe+oo
o.ooooooE+oo
QW
03000001900
Qme
QW
o.oooooos+oo
QW
QWEd-Q)
QWEd-m
QW
0W
0W

Figure 41: Results of Four Stage Interconnection Pattern with Sigmoid Neuron.

125

ML

 

 

SIGNAL NAMES I
I
(NS? I K001113) K001'04) mums.) K001116) K001117) K001118)
O |I aoooooom OW Q W OW 000000084» QW
4 I LWEHX) 1W Lamooomoo LW moooooewo LWEW
14 I o.oooooo£+oo QW Q ooooooa+oo o.oooooaa+oo QW o.oooooos+oo
24 I 1900000300 LW SW01 LW 8m01 8m01
34 I QWE-rm QW Q 000000300 o.ooooouE+oo QW o.oooooos+oo
44 I 1WOI 1000000501 aoooooomoo 14000000300 o.ooooooa+oo QW
54 I QWEHX) o.owoooa+oo Q mm mamas-01 o.ooooooE+oo QW
64 I QWEi-W 14500000501 Q oooooua+oo LW QWE‘KX) QW
74 I o.oooooo£+oo QW ammo 1.oooooas+oo QW 00000001900
84 I o.oooooo£+oo QW Q oooooomoo Looooooe+oo QW 0900000300
94 I QWEHX) QWEd-Q) Q oooooomoo Loooooaa+oo 0.000on+00 0900000900
104 I QW 09000001900 Q 000mm LW o.oooooo£+oo QW
114 I QWEHX) 09000001300 0. oooooaawo 1.1!er(1) QWEi-(X) o.ooooooa+oo
124 I o.ooooooa+oo o.oooooor-:+oo o.oooooon+oo monomwo omooooewo QWE‘I‘Q)
134 I o.oooooos+oo 0W W LW onooooomoo QW
11MB L SIGNAL NAMES I
I
(NSI) I K001119) K001120) K00'I121) KW'1122) 1:00 1123) K001124)
o I comm-mo OW Q W OW omoooa-zwo QW
4 I‘ momma-co Imoooomoo LW 1W 1W URINE-rm
14 I QWE-I-w o.oooooos+oo Q mm ammo (1000000300 09000005400
24 I 1 oooooos+oo LW 8m01 W01 Loooooomoo 1 .W
34 I Q WEI-(X) QW omomoswo Q W o.ooooooa+oo Q 0000001900
44 I 1500000501 1500000501 QW QWEM LW 501111501
54 I 0. WEI-(X) 0.0000me QW Q oooooomoo o.ooooooa+oo Q ooooooswo
64 I OWE-(X) momma-01 0900000900 QWE‘FN LWM 1500000301
74 I Q oooooomoo QW common-mo 0W W01 QW
84 I 04000000900 QW o.ooooooa+oo 0mm Loooooos+oo o.oooooo£+oo
94 I o.oooooo£+oo QWEM o.ooooooa+oo o.oomooa+oo 19000005400 ammo
104 I QWE-I-m ammo QW QW 1.00000054-00 QWEd-m
114 I QW 09000005400 o.oooooa-:+oo o.mooooE+oo 1.ooooom~:+oo o.ooooooE+oo
124 I 0.000000% 0W ammo QW 19000001900 QWEd-Q)
134 I ammo common 0W QW wooooomoo o.ooooooa+oo

 

 

Figure 41. (continued).

126

 

11MB E SIGNAL NAMES I
I
(NS) I K001125) K001126) K001127) K001128) K001129) K001’(30)

(100000015400 o.oooooos+oo QW omoooom o.oooooua+oo QW
LWEM LWE-I-m LW LW Loooooomoo l.QXXmE-I-(XI
o.ooooooa+oo QW omoooomoo omoooomoo QW Q WK!)
LWEKX) LW Looooooewo LWM LOWEM 1 “BOWEN
o.oooooos+oo QW omoooo£+oo QW‘IQ) QWEd-Q) Q ooooooB+oo
1 .0000:an 1.1!!!an40) LW-tm Looooooa+oo 8.5WE01 l.QXXXXIE-rm
Q WEI-(XI QW 00000005410 omooouawo QW o.oooooo£+oo
1 .oooooomoo LWOE-o-m 1900000900 Looooomwo asooooos-m LWEHXI
Q oooooo£+oo o.oooooos+oo o.ooooooe+oo o.oooooos+oo QW o.oooooora+oo
o.oooooos+oo QW-Ol QW QW o.oooooo£+oo Q WEI-(XI
QWEAX) 8W1 o.ooooom+oo aoooooomoo QWEKX) Q oooooos+oo
o.ooooooa+oo 8m01 QW Q WEI-(X) QW Q WEI-(X)
o.ooooooe+oo 8m01 QW Q WEN!) QW QWE‘o-Q)
QW M01 QW 0W 0300000900 0900000900
QW W01 QW OW QW QW

E§E§££§sssss§*°—

 

 

-3_
E
E

048') I K001131) K001132) K001 (33) K001134) K0010!) K001136)

0 I QW 00000005400 QW ammo 0mm QW
4 I MIME-Pm 1mm LW momma-co 1mm Lomooomoo
14 I o. oooooos+oo QW o.oooooo£+oo o.oomo(£+oo QW o.oooooox~:+oo
24 I 140000001900 LW ”KIM-0m 1W 1000000300 Looooooewo
34 I o.ooooooE+oo QW o.ooooooe+oo 000mm o.oooooo£+oo 0900000900
44 I immommo hm“) LW'I-m Looooamoo Loooooomoo Looooooe+oo
54 I QWEd-(XI o.ooooooe+oo 00000001900 09000001900 QW 00000001900
64 I LIXXXXDE-o-w Loooooos+oo LW moooooaoo LW 1900000900
74 I QWE'KXI QW 090000013400 omoooa-zwo Q oooooos+oo oooooooewo
84 I QWEKX) QW monsoon-01 QW 5.m01 1MB“)
94 I 09000005400 QWM 09000001900 00000001900 QW o .ooooooewo
104 I o. WEN!) OW OW QW 0900000900 1 .ooooooaoo
114 I Q (111000544X) QW QW o.ooooool=.+oo o.oooooo£+oo 1100000304!)
:3 : Q WEI-(l) 0mm 0W QW o.ooooooa+oo 1 .WEHX)

Q oooooos+oo OW OM44!) common 0mm LWAD

Figure 41. (continued).

127

“MEL

 

 

SIGNAL NAMES I
I
(NS? I K001137) K001138) K001139) KOUT(40) K001141) K001 (42)
O I aoooooow omoommoo Q W OW omoooa-zwo aoooooom-oo
4‘ I 1MB“) LWE-o-m 1 mom-mo LW mooooomoo LWEM
14 I o.oooooos+oo 00000005400 Q 000000900 onoooocmoo QWEHX) QWE-I-(X)
24 I LWEM LW 10000005410 Loooooaawo LW LWEM
34 I omoooomoo QW omoooomoo cocoon-mo 0.00000054-00 QWEAX)
44 I 1mm LW LW-tm 1W LW LWM
54 I o.oooooo£+oo QW 0mm onoooooaoo Q W o.oooooo£+oo
64 I mooooom-oo Lomoooewo 13000001900 LWEM 1 .ooooooE+oo moooooem
74 I ooooooo£+oo QW QWEHX) omooormoo Q ooooooaooo 09000001900
84 I moooooaoo hm“) LWE-o-m mooooosmo LW 1.oooooos+oo
94 I omoooos+oo Q WEI-Q) o.ooooooa+oo o.oooooaa+oo QWEHXI QWEM
104 I 1500000301 1500mm o.oooooo£+oo 1.000000£+oo QW S.W)E01
114 I o.oooooos+oo OW ammo 1W1 QW QWE-om
124 I QWEHXI onooooomoo omoooomoo LW omoooosm QWEM
134 I OWN omoooos-roo 0W LW OW o.ooooooa+oo
11MB L SIGNAL NAMES I
I
(143.) I K001143) K001144) K0011“) K001146) K001147) K001148)
O I QW OW Q W OW omoootmoo QW
4 I momma-00 moommo LW imooooam mooooomoo l.QXXXIOEm
14 I QWEW ammo 00000001900 ammo 00000005400 09000001900
24 I LWEHD LWW LWIEW 1mm LW Loooooos+oo
34 I QWEKX) Q W ammo omooooawo QW 0000000900
44 I LWEM 1 .W Loooooomoo 1.000001:st LW 1900000900
54 I QWE'o-Q) o.oooooor-:+oo omoooomoo 0.000000st 0.0000005+oo 0000000900
64 I l.QXIOIDEé-(XI LW LWKX) Looooooaoo LIXXXXXJEHX) Loooooos+oo
74 I o.ooooooa+oo QW 00000005400 09000001900 QW QWEAX)
84 I 1WE+M LOWE-Q) LWEM LW-o-m momma-+00 1.oooooos+oo
94 I o.oooooos+oo QWEd-(D QWEM o.oooooo£+oo QW QWE-I-(X)
104 I Loooooos+oo o.ooomoe+oo omooowwo o.oooooomoo o.oooooos+oo O.(IXIIOOE+(XI
114 I S.(XXXXIOE01 04000000900 QWE-I-(X) 00000001900 QW QWEW
124 I Looooooa+oo OWE-rm omooooswo (1000000900 Q oooooomoo 0300000900
134 I hm 0W 0110000er aoooooom-oo 00000001900 o.oooooos+oo

 

 

Figure 41. (continued).

128

B
.5

 

“MEL SIGNALNAMES I
I

018) I K0010) Km K00113) K00114) K00115) K00116)

OW QW 0000mm QW QW
moooowwo 1900mm LW MINNIE-om LW imoooomoo
1.ooooooB+oo ammo QW QWEd-(X) o.ooooooa+oo QW
1300000300 omoootmoo amen QWE-HXI o.oooooos+oo QW
Looooooa+oo omoooaawo OW QW omoooomoo QW
LW onooootmoo 0W QW omoooomoo QW
onooooa-zwo QW QW 0mm QW
1.000000% OW 0W QW o.oooooon+oo QW
LW omoooaa+oo QW QW ooooooomoo omoooomoo
LW omooooaoo o.oowooa+oo 0.000000% omooooe+00 o.oooooos+oo
LW omooormoo OW QW QWEAX) QW
LIXXXXDEM QW ammo QWE-I-Q) QMKX) QWEHX)
Loooooos+oo o.oomoor-:+oo onoooooewo 0W onooootmoo 000mm
mooooos+oo QW QW ooooooo£+oo ooooooomoo QW
1000000900 QW Qm‘tw QWE-I-m 0000mm QW
moooooswo QW omoootmoo QW uncommon 00000001900
moooooaoo o.ooooooa+oo o.ooooom+oo o.ooooooe+oo ammo QWE-rm
. . . o.oooooa-:+oo QW
1.oooooo£+oo QW o.ooooooE+oo QWE-o-m QWi-(XI o.ooooooa+oo
1.oooooo£+oo QW onoooou-zwo 0W4“) o.oooooaa+oo 0W
1.0000003+00 QW 0000000300 00000005400 00000005400 QW
10000005300 QW 0000005400 0W o.oooooaa+oo QW
1mm W 0mm OW OW 0W

§§§§§§E§§E§flnurrzm
é S
E
§
§
3
é

8
a.

SIGNAL NAMES I
K00119) MIG) “1101) K001112)

-32.;
E I
I
I
I

W
'1.oooooo£+oo moooooewo LW-I-(D 1mm 1mm LW

o.oooooo£+oo QW ammo 09000002400 QW omoooomoo

on.
fi

I
24 I o.oooooon+oo QW Looooooam o.ooooooa+oo QW OW
34 I onooooo£+oo QW moooooawo omoooaa+oo cameos-.00 000000015400
44 I QWE-I-m QW Loooooos+oo o.ooooooa+oo aoooooomoo 09000008410
54 I Qme o.ooooooe+oo MIME-om o.oooooos+oo QW 09000001900
64 I QWE-I-(X) o.owoooa+oo ”KNEW o.oooooos+oo QWEM 0000000900
74 I o.oooooos+oo QW woooooa+oo o.ooooom=.+oo QW 03000005410
84 I omoooomoo QW LW o.oooooos+oo QW QWEKX)
94 I QWEd-Q) QW LWEKD o.ooooooa+oo aooooooawo o.ooooooa+oo
104 I 00000001900 0 oooooomoo 1 oooooomoo ammo 0 000000900 0 WEI-Q)
114 I O oooooomoo 0 mm 1 0000001900 QW o WEI-(X) 0 W441)
124 I OWE-(XI 0W 10000001900 QW oooooooe+oo OW
134 I o MOE-Kl) 0 0000001900 1 oooooomoo QWEHX) 0 0000001900 0 W
144 I o M00054“) 0 oooooomoo moooomwo QW 0 0000001900 0 W
154 I 0 0000005400 omoooomoo 1 W QW o oooooomoo 0 W
164 I o oooooos+oo 0 000000900 woooooawo QW 0 000000900 0 ooooooawo
174 I QW 000000013400 1 mm QW o oooooomoo o W
184 I o oooooom-oo 0 mm 1 comm o.oooooos+oo O cocoons-+00 0 000000300
% : O ooooooa+oo omoooozwo 1 ooooommoo O W 0 000mm 0 W
I

Figure 42: Results of Two Stage Interconnection Pattern and Sub-optimal Path.

129

 

 

 

 

Figure 42. (continued).

130

11MB I SIGNAL NAMES I
I
048') I K00103) KGJ'I'04) K00105) M16) KGJ'I‘07) K00108)
o I QW 0W Q W OW o.ooooooa+oo QW
4 I 1mm Loooooomoo 1 .W Loooooozm 1mm MINNIE-(X)
14 I QWd-IXI QW Q ooooooe+oo o.ooooooE+oo commas-too o.oooooo£+oo
24 I Loooooomoo LW LW moooowwo LWOE-o-m LWE-om
34 I OWN!) QW onooooomoo omoooomoo QW o.ooooooe+oo
44 I QWE-tm QW QW 1.ooooom+oo aoooooomoo o.oooom1=.+oo
54 I omoooomoo commas-too omoooomoo MIME-0m QW o.ooooooa+oo
64 I o.oooooos+oo QW o.oooooos+oo HINGE-om QW o.ooooooe+oo
74 I OWEN!) QW QW woooooawo QWM QWE-I-(X)
84 I QWE-I-(X) QW QWM mooooaawo o.oooooos+oo o.ooooooa+oo
94 I o.ooooooE+oo QW o.ooomoa+oo LW QW QWE-om
104 I QW o.oooooor=.+oo uncommon LW 09000001900 o.ooooooe+oo
114 I QW o.ooooooa+oo QW LW o.oooooos+oo o.oooooos+oo
124 I QW 00000005410 onooootmoo LWE-I-Q) o.oooooos+oo QWEAD
134 I QW o.oooooos+oo omooooswo LW QWEM QW
144 I o.oooooos+oo OWE-IQ) o.oooooor-:+oo LW QCXXIOQJEI-m QW
134 I QW o.ooooooa+oo QWi-N LW QWE-I-(X) QW
164 I QW 0mm QW 11!me oooooooewo QWEW
174 I QW 0W 0.0mm LW omoooomoo QWE-tm
184 I QW omooooa+oo o.ooooom-:+oo LW QWE-KXI 00000001900
194 I QW omoooomoo 0900mm kW o.oooooo£+oo QWXIEH”
204 I QW QWEM o.oooooa=.+oo LWM o.ooooooa+oo o.oooooo.=.+oo
214 I QW OW uncommon LW OW QW
111:3 I SIGNAL NAMES I
(148.) I K00109) W) K001 (21) K001(22) K001123) K001 (74)
O I aoooooom-oo OW Q W 0W OWN QW
4 I. LW moooooawo 1 .W LW uncommon LWHX)
14 I QWEM QW omoooomoo o.oooooa=.+oo QW o.oooooo£+oo
24 I Loooooomoo LW Looooooaoo LW'I-(XI 1.0000005+oo LIXXXXXIEM
34 I 09000001900 QWE-I-Q) 00000001900 o.ooooooa+oo QW o.ooooooe+oo
44 I LUMBER!) LWE-I-m 1.ooooooa+oo 1.oooooo£+oo LW LIXXXXDE'I-(X)
54 I o.oooooor-:+oo o.oooooos+oo o.oooooos+oo omooooewo o.oooooos+oo o.ooooom-:+oo
64 I omoooomoo LW o.oooooos+oo omoooaawo LW Q oooowswo
74 I o.oooooos+oo o.oooooos+oo QWE'KXI o.ooooooa+oo QW Q ooooooa+oo
84 I o.ooooooa+oo LW o.oooooo.=.+oo o.oooooa-:+oo LW Q WEI-(XI
94 I omoommoo o.ooooooa+oo OWEN!) o.oooooa-:+oo QW QWEM
104 I QW LW-rm o.oooooua+oo QW l.QXIQDEM QW
114 I 00000001900 0mm omooooaoo QW QWEKX) o.oooooos+oo
124 I QW mooooomoo o.ooooo(£+oo QW 1110000013410 QW
134 I QW 00000001900 uncommon o.oooooma+oo QW QWEHD
144 I QW imoooomoo o.ooooou£+oo o.oooooos+oo 1.ooooooa+oo QW
154 I 000000015400 omoooomoo o.oooooaa+oo QW o.oooooor-:+oo o.oooooos+oo
164 l o.ooooooe+oo 1mm . 0900000900 aoooooomoo 1.a)ooooa+oo QW
174 I QW o.oooooon+oo ammo o.ooooooa+oo o.ooooma+oo 0.000000154-00
184 I QW mooooomoo o.oooooaa+oo QW LWEM o.ooooooa+oo
194 I QW o.ooooooa+oo o.oooooos+oo QW QWEd-(X) QWOE-I-(D
”4 I QW 11100000300 omooooswo aooooooB+oo momma-+00 QW
214 I QW uncommon 0.0mm QW Q 000mm

o.oooooos+oo

 

 

 

 

Figure 42. (continued).

131

TIME I SIGNAL NAMES '
I
(NS) I K001125) K001126) K001 (27) K00'1128) K00'1129) K001 (30)
' A
0 I QWEHX) OW QW 00000008400 omoooamoo W
4 I Loooooomoo Looooooa+oo LW 1W imooooawo LW
14 I Q WEI-(X) o.ooooooE+oo o.ooooooa+oo o.ooooooa+oo 001000314!) o.oooooos+oo
24 I 1.oooooo£+oo Looooooa+oo LWE-o-m LWM 10000005400 Loooooomoo
34 I Q WEI-(X) QW o.oooooos+oo omooooawo QW o.ooooooE+oo
44 I 1.ooooooa+oo LW 1.oooooos+oo LW-Im Looooooa+oo 1.00000084-00
54 I 090000015400 QW o.oooooos+oo 0W4“) aoooooomoo o.oooomE+oo
64 I 1.oooooo£+oo LW LW LWM l.QDOIXIE-I-(X) LWM
74 I o.oooomE+oo QW 09000003410 o.ooooom=.+oo QW onooooomoo
84 I Loooooomoo LWEHX) Loooooomoo moooooewo Locoooomoo LWE‘I-Q)
94 I QWEM o.oooooo£+oo o.ooooooa+oo o.ooooooe-oo QW o.ooooooa+oo
104 I LIXXXXIOEKXI Looooooem 1.00000(E+oo 190000015400 Looooooa+oo Loooooos+oo
114 I QWE-I-Q) omoooomoo o.ooooow+oo ooooooomoo o.ooooooE+oo QWE-I-Q)
124 I LONGER!) Loooooomoo LW‘KD 1.0100084“) Loooooomoo Looooooa+oo
134 I o.ooooooa+oo o.ooooooa+oo 0900000900 QW QWEi-m QWEKX)
144 I l.QXXDOE-tm Looooooa+oo LWM LW 1 .ooooooawo Lowoooewo
154 I QW o.ooooooa+oo ammo QW Q WEAK) omooooewo
164 I 1.111100%“) 1000000300 LWM 1.oooooos+oo LIXXXXDEAI) mmoomoo
174 I 0.1110008“) onoooooaoo o.oooooo£+oo o.ooooooa+oo o.ooooooE+on I‘. I: W. memo
184 I Looooooewo LWE‘IQI 1.ooooom+oo Loooooos+oo 1.oooooos+oo IUJIKXIOE‘I-(D
194 I QIXXXIOOEAX) 09000005400 0000000500 QW Q 00000015400 ouxmosmo
204 I Looooooa+oo Loooooomoo 1 .ooooooswo LWENX) 1 .oooooomoo l.QXXXXJEm
214 I woman-+00 o.oooooos+oo 0mm QW 0000000900 o.ooooooa+oo
115:5 I SIGNAL NAMES I
018') I K001131) m2) K001 (33) K001(34) K001135) K001 (36)
O I o.oooooor=.+oo 0W Q W OW onooootmoo QM“)
4 I- LWE'o-w Loooooos+oo 1 .W 1W 1mm LW
14 I 040000001900 QWEHD Q oooooom-oo QWM QW 00000001900
24 I 1.0001054“) LWEKD 1110mm moooomwo LW LWE‘HX)
34 I o .oooooomoo 0.000000% o.oooooos+oo o.ooooooa+oo QW o.oooooos+oo
44 I 1 .oooooos+oo LW 1.ooooool-:+oo LWM Loooooomoo LWE‘KX)
54 I o.ooooooa+oo o.ooooooa+oo Q ooooooewo QWHX) QWE-I-(X) QWEI-(XI
64 I 1.oooooo£+oo LIXNXJEHD 1.000000£+oo 1.oooooos+oo LOWE-rm 19000001900
74 I 0100003544!) QW Q «momma o.ooooooa+oo 0.00000024-00 QWEM
84 I Loooooomoo LW 1.ooooom=.+oo Loooooomoo LW Loooooos+oo
94 I 0900000900 0.00000024-00 QWEM QWKX) QW O.QXXJQ)E+(XI
104 I Looooooa+oo LWIE-I-(XI LWKII LWW 1.0000(an l.QXXIOOE-I-(X)
114 I o.oooooos+oo o.ooomoa+oo o.oooooo!~:+oo QWW o.oooomE+oo QW
124 I 1111000844!) Looooooawo LW-tm LW 1.000000st 1.00000054-00
134 I Q W ammo o.oooooos+oo QWEd-(XI o.oooooo£+oo o.oooooos+oo
144 I 1 .ooooooewo LW'IQ) LIXXXXJIEHX) 10000001900 1.ooooooE+oo l.QIOOOOE-I-Q)
154 I Q oooooomoo o.oooooos+oo o.oooooma+oo QW omoooomoo 01100001544!)
164 I 1 .ooooooa-oo Loooooomoo 1.oooooaa+oo LW 1.oooooos+oo 1.moooos+oo
174 I o.oooooos+oo OW omooouawo aoooooomoo QOIXIOQJEM QQIOOQIEAD
184 I LW LWOEM 1.oooooa-:+oo l.QXDQIEM LWE‘HX) l.QXIIXIOEHD
194 I o.ooooooa+oo o.oooooos+oo QW-I-Q) QW o.ooooooE+oo Q 000000300
204 I Loooooomoo 1 ooooooswo mooooomoo Loooooosm 1 WEI-(X) 1 .ooooooE+oo
214 I QW Q 000mm o.ooooom+oo QW omoooomoo QW

1'qu
(NSH

SIGNALNAMBE I

ROUND) K0011“)

OW
1W
o0ooooas+00
1000000500
0W
10000005+00
o00000mwo
1000000300
0000000900
1000000500

K0010?) mm)
0000000900
1000000900
00000005000
10000001900
000000015900
1000000500
0000000300
1000000500
00000001900
1000000500
0000000500
10000001300
00000005+00
1000000900
00000001900
10000005400
00000002200
10000005+oo
0.000000£+00
10000005+00
0000000500
10000001900
00000001900

QW
100000015400
0000000900
LWEM
0.0000(05+oo
10000005300
00000001300
10000001300
0000000£+00
LWEW
QWEd-Q)
10000005400
00000005400
1000000£+00
QWOEI-Q)
LWOE-I-QI
0.000000£+00
LW
000000012400

:2!!§2!t!§i*°_

§E§§§§§§“

_§-_________

8
&

SIGNAL NAMES 5
K001 (45) W46) KOUTQ'I)

:3

I zooms) zoo-rm) Karma)

 

---3

fixixtrfizifriiitffireo

g—l—Iup—I—o—ouu—
—-—--—---

23
a.

0000000£+00
1.0000001=.+00
0000000200
1 .QXXXDEM
0.0000001-:+00
LWEd-(X)
QWEKX)
1 0000005+00
QOIXJOIDEHX)
1000000900
OWE-(X)
100000013+00
00000001-:+00
100000013+00
o000000£+00
1.0100084!)
0.0000001-:+00
1000000900
0000000000
1000000500
0000000300
100000015400
OW

00000002400
1000000£+00
QWE-I-Q)
11101084411
00000005410

00000001900

0mm

0W
LW
00000001900
100000015200
00000001900

0mm

0W
1.00000013+00
o00000¢na+oo
LWIEHXI
0000000200
1000mm
00000005200
100000ma+00

Figure 42. (continued).

132

10000001900

000mm QW

LW
0000000300
1 000000500
QWEM
10000001900
o0000001=.+oo
LIXXXXXIE'KX)
0000000900
1000000500
00000005+00
LWEHX)
o.ooooooa+00
l.QXXXXIE'I-Q)
0000000300
1 0000005+00
0000000300
LW
0000000500
LWEKD
0000000£+00
LW
QW

111:5L
NH 3001(1)
I

o I 00000001500
1mm
LW
10000005200
LWEHX)
1000000300
LWEHXI
LWE-I-(D
10000005200
10000001900
LWOE-KX)
1000000£+00
1.0000001-:+00
1000000200
moooowwo

§§E§rrrrrtrrrg

-%

3.2
I

if

§§E§553255555‘°
é

Tgff

Km

Q W
1 .W
Q ooooooaoo
0000000900
00000008200
00000001500
0mm
0000000900
0000000300
00000003400

K0010)

QW
10000002+00
0000000500
0000000900

SIGNAL NAMES
K00114)

SIGNAL NAMES

OW

K00118) K00119) K00100)

Q W
1 .W
00000001900
10000001500
1.101103%
LWM
1000000300
LW-rw
100000015400
100000015400
10000001500
LWXEKX)
10000003+oo
100000t£+00
100mm

OW
1W
0000000£+00
QW
QWd-(D
0000000£+00
00000005400
0000000900
0000000900
o00000aa+00
000000(£+00

300115)

0. W
1 000000300
Q 000000500
Q 000000800
Q 000000900
Q W
Q WEI-(X)
Q W
Q 000000st
Q 000000900
00000005400
0000000£+00
0000000900
0000000500
OW

K00101)

K0016)

K00102)

Figure 43: Results of Three Stage Interconnection Pattern and Sub-optimal Path.

133

11MEI

 

 

 

 

Figure 43. (continued).

134

SIGNAL NAMEE I
I
048') I K00103) K00104) K00105) K00106) K00107) K00108)
O 'I o000000£+00 0000000500 Q W OW 0000000300 QW
4 I LNOOEW 10000005400 1 000000300 1W 100000tm00 1000000300
14 I 00000002400 0000000£+00 OWN!) 000mm QW QWEHXI
24 I LWE-o-(D LW 00000005300 100000aa+0o QWOM 0 000000500
34 I OWEN!) 0000000500 0000000900 0000000500 0000000500 0 .(XXXXXIE-I-(X)
44 I 0000000900 00000005400 00000003400 10100013200 QWEHX) 00000001900
54 I OWE-04X) QW 0000000900 LWHD QWEd-(XI QWE-HX)
64 I 00000002400 QW 00000001900 1.00000u-:+00 0000000900 QWEd-(X)
74 I 0000000900 QW 0000000500 100mm QW 0.0000001=.+00
84 I QWEHXI QW 0000000500 10000005400 QW 000mm
94 I 0000000500 QW 0000000300 1000000500 QW 0.0000001-:+00
104 I 0000000900 0000000500 00000005400 100000012+00 QWEM 0.“!!me
114 I 00000005+00 QWi-(X) 000000th LW 0000000£+00 0000000500
124 I 0000000£+00 0000110500 0000000500 LW 00000001500 QW
134 I 00000005000 OW 0W LW 0000000900 QW
11MB L SIGNAL NAMEE I
I
CNS.) I K00109) Km K001121) K001122) K001123) K001 (24)
O I- 0.000000£+00 0mm OW OW 0mm comooomoo
4 I 1000000300 LWXIEM LW 1.000000£+00 100mm LW
14 I QWEM 000000013+00 o00000015+00 00000001900 Q W 0000000500
24 I LWEM LW LWOE-o-m 1000000900 1 .W 1000000500
34 I QWEW 0000000500 00000001900 00000005400 QWEHXI Q 000000st
44 I OWE-(l) 1 .W 1000000300 1000000500 1W 10000005000
54 I 00000001900 Q W 0000000900 0.00000(£+00 QW Q 0000001900
64 I 00000001900 Q WORM 00000001900 00000001900 LWOEHD 0000000500
74 I OWEN!) Q W 00000001900 OWN!) LW 00000008+00
84 I QWEHX) 0000000900 o0000005+00 o0ooooos+oo LW 00000001900
94 I QWEM QW 000mm 0000000900 LW QW‘KX)
104 I QWOE'KXI Q 000000300 QWKX) QW LW'I-(X) 0.1!!!me
114 I QIXXXIOOEM Q 000000500 0000000500 QW 10ooooos+oo 000000013+00
124 I QIXXXIOOEHX) 0000000500 0.00000(£+00 QW 1000000900 QWEHD
134 I 0000000£+00 o0000005+00 00000001900 QW 1000mm QW

MI

§§E§rritxtrgztg

-§

ffiflifitffiitg

§§§§

I
(NSI): mums

aooooooswo
LW-rm
09000:an
19000005400
01100000300
1.ooooooB+oo
o.oooooo£+oo
Locomoaoo
omooooswo
00000002400

Figure 43. (continued).

Karma)

QW
LWEHXI
QWEM
l.QXXXXIB-I-w
0900000900
1.“!!!an
o.oooooo£+oo
lm-KX)
0000000300
09000005400
ammo
o.ooooooa+oo

o.ooooooa+oo
QW

 

 

TIME I SIGNAL NAMES I
I .
015') I KOUI‘(37) KOUTG8) KOUTG9) KOUTGO) KOUT(41) KOU'1'(42)
0 I W 0W QW 0W 0W QW
4 I 1.ooooooa+oo 190000013400 LW 1W Loooooaawo Lomoooswo
14 I QWEM 00000005410 0W 0W QW 0000mm
24 I 1.ooooooE+oo LWM 19000005410 mooooamo LW LWM
34 I o.ooooooa+oo QW omoooomoo omooomwo QW o.ooooooa+oo
44 I Loooooomoo LW LWEd-(l) 1900000300 MINNIE-04X) LWB-rm
54 I onoooooswo QW 0W o.oooooaa+oo QW o.ooooooa+oo
64 I LWEW LW LWEHX) LW-tm LW Looooooswo
74 I QWEHX) QW-o-Q) o.oooooos+oo omooooaoo QW o.ooooooa+oo
84 I 10000002400 LW-O-(D Loooooomoo LW LW 1mm
94 I QWEM o.oooooos+oo o.ooooooB+oo o.oooooos+oo «0000001300 00000001900
104 I LIXDOOOE-o-m moooooam omoooomoo LW omoooomoo 11100000900
114 I o.oomooB+oo o.ooooooa+oo 0W QW 00000002410 ammo
124 I noooooomoo Q ooooooe+oo omooooawo LW omooooe+oo QWEM
134 I 09000001900 omooooawo 0W LW 0W 0W
TIME I SIGNAL NAMES I
I
018') _ I KOUK43) KOU'I'(44) KOUT(45) KOUTGQ KOU'I‘(47) KOU'I‘(48)
0 I aoooooo£+oo onooooomoo 030000013400 0W OHM-KI) QW
4 I 1.WE+€X) 1mm momma moooooe+oo Loooooaswo LW
14 I QWEM QW o.ooooooa+oo omooooawo 0.000000% 0.000000%»
24 I Looooooa+oo LW-I-m mooooomoo 19000005400 momma-+00 LW-tm
34 I o.oooooox~:+oo o.oooooo£+oo omoooomoo o.oooooas+oo QW o.ooooooa+oo
44 I MINER!) Lowoooswo 1900000300 Looooowwo LWM 1.oooooo£+oo
54 I QWEM QW 0411000001900 0900000500 00000001900 Q WEI-(II
64 I l.QXXXXJE-om LW Looooooa+oo 1.ooooom+oo LW 19000001900
74 I o.oooooor.~:+oo QW omoooomoo o.oooooaa+oo Q W Q W44!)
84 I 1.ooooooa+oo LW 1mm Looooom+oo Locooooswo 10000005400
94 I QWE-o-(X) 0300000300 OWEN!) o.oooooas+oo Q 0000005300 omoooomoo
104 I 1.mOOE+00 1900000900 moooocmoo 1.000000£+oo 1.oooooos+oo Looooooa+oo
114 I QWOOOBd-Q) Q W44!) o.ooooooB+oo o.ooooooE+oo QWEM o.oooooot-:+oo
174 I 1.00000054-00 Q oooooomoo onooooms+oo noooooos+oo 0900000300 o.oooooos+oo
134 I 1.oooooos+oo Q WEI-(X) woman-r00 aoooooomoo 0W QW

 

 

Figure 43. (continued).

136

 

KOUTO)

0 I omooocmoo
4 I Looooooewo
LW
LWOW
1.000000£+oo
1.oooooo£+oo
Looooooa+oo
Loowoomoo
LWEd-m
LW
l.QWJEe-N

ffififlfitfifii

1r-

-§

EIIIIE é

ffifiittfifiie°

ffii‘i!t!§i*°

3-2 _________

1-

onooooo£+oo
QWE-rm
0000000900
000000013410
onooooomoo

KOUT(13)

o.ooooool~:+oo
19000001900
QWEM
04000000900
oooooooawo
QWE-KXI
o.ooooooa+oo
omooooswo
omooooswo

Km

QW
LW
o.ooooorn+oo
omoootmoo
00000005410
ooooootmoo
0W
omoooomoo
o.oooooaa+oo
0W
0.0mm

KOUTG)

300119)

SIGNALNAMES

KOUTG) KOUT(4)

MAI-NAMES

ammo
LW-o-m

KGJ'IUO)

K0076) 3007(6)

o.oooooon+oo omoooomoo
LW moooooaoo
ooooooomoo o.ooooooa+oo
OWE-o4!)
QWEAD
0900000900
o.oooooomoo
00000001900
03000001900
QW
0W

KOUTO 1) KOUT-f 11’. 5

0mm
LW
09000001900

SIGNALNAMBS

0W
LW
o.oooooaa+oo
04000001300
omooormoo
omoooomoo
0W
0900mm
0000001300
0W
0W

KOUT(17)

W
LW
o.oooooma+oo
o.oooooo£+oo
o.oooooos+oo
omoooomoo
ooooooos+oo
0000000900
onooooomoo
onooooomoo
0W

KM“)

ammo QW

1mm
oooooooawo

LW
09000005410
LWE-rm
09000008400
ooooooomoo
QWE-I-m
00000005400
omooooawo
0000000900
0mm

Figure 44: Results of Four Stage Interconnection Pattern and Sub-optimal Path.

137

2-3-

ifflgfitf'fi'i‘O

fifflflfit‘fifiiecw

___g_g rrrrrtrrzt

...3.3

P

1

KOU'1'(19)
QW

I

ooooooomoo
1.oooooo£+oo
QWE'KD
ooooooomoo
o.oooooor-:+oo
0.0000me
QOOOOQIEM
09000003400
0900000300

W)

KOU'1‘(21)

Loooomoo
00000005410
mooooomoo
00000005400
0W
omooooawo
ooooooomoo
o.oooooo£+oo
0.0mm
0W

 

1-

1..

onoooooawo

KOUT(31)

aooooooaoo
1900000300
0900000900
Loooooomoo
o.ooooooa+oo
1000000300
o.oooan5+oo
o.ooooooE+oo
o.oooooo£+oo
o.oooooos+oo
omooooaoo

KOUTGS)
o.oooooos+oo
omoooomoo
moooooawo
o.oooooo.=.+oo
000mm
000mm
omoooomoo

omoooomoo

SIGNALNAMBS

SIGNALNAMBS

xoumz) M)

1mm
0mm
1mm
o.ooooma+oo
omooooswo
ooooooomoo
0mm
00000003400
ooooooamo
mm

Figure 44. (continued).

138

0mm
1mm
QW
Loooooomoo
QW
LW
woooooaoo
LWXIEKX)
LWHXI
1W
LW

xomm)

QW
Loooooos+oo
o.oooooos+oo
mooooomoo
o.oooooot-:+oo
o.oooom+oo
o.oooooox~:+oo
omoooo£+oo
o.oooom1=.+oo
04000000900
o.oooooo£+oo

onooooomoo

LWW

TIME!

I
(NS') I XOUT(37) KOU'I'OS) £00709)

(1000000900
13000001240
OWEN!)
LWE-om
ooooooomoo
LWE'HD
omoooomoo
moooooswo
OWEN!)
00000002410
o.ooooom-:+oo

runner»-

3-;

I trauma)

onmooswo
1W
09000001300
LWHXI
amour-2m
mooooomoo
QWi-(X)
1.oooooos+oo
OWE-04!)
moooooaoo
1.oooooo£+oo

runner”

QW
1 .W
o.ooooooa+oo
moooooawo
ammo
wooooomoo
o.ooooooa+oo
1mm
o.ooooooa+oo
0.0mm
0W

SIGNALNAMBS

W40)

0W
1W
ammo
1W
0mm
1mm
ooooooa-za-oo
1mm
o.oooooa~:+oo
moooooaoo
1W

Figure 44. (continued).

139

KM“)

0mm
Looooamoo

xou'mz)

QW
LW
00000001900
LWE‘KD
o.ooooooa+oo
Loooooom-oo
o.ooooooe+oo
Looooooa+oo
omooooewo
omoooo£+oo
0W

References

140

 

References

[1] R. Jurgen, ”The Specialties”, IEEE Spectrum, January 1991, p. 79.

[2] R. Lippmann, ”An Introduction To Computing With Neural Nets”, IEEE ASSP
Magazine, April 1987, pp. 4-22.

[3] F. Salam, ”A Tutorial Workshop On Neural Nets And Their Engineering Imple-

mentations”, 31 st Midwest Symposium 0n Circuits And Systems.

[4] M. Kennedy and L. Chua, ”Neural Networks For Nonlinear Programming”, IEEE'
Transactions on Circuits and Systems, Vol. 35, N o. 5, pp. 554-562.

[5] M. Hanes, S. Ahslt, K. Mirza, and D. Orin, ”A Neural Network Interface to the
DIGITS Grasping System”, ICJNN Conference on Neural Networks, 1990, Vol.
III, pp. 343-348.

[6] S. Wang and H. Yeh, ”Self-Adaptive Neural Architectures for Control Applica-
tions”, ICJNN Conference on Neural Networks, 1990, Vol. III, pp. 309-313.

[7] W. McCulloch and W. Pitts, ”A Logical Calculus of the Ideas Imminent in
Nervous Activity”, Bulletin of Mathematical Biophysics, Vol. 5, pp. 115-133,
1943

[8] J. Vidal, J. Pemberton, and J. Goodwin, ”Implementing Neural Nets with Pro-
grammable Logic”, IEE'E 1st Conference on Neural Networks, 1987, Vol. III, pp.
539-545.

[9] J. Hopfield, ”Neural Networks and Physical Systems with Emergent Collective
Computational Abilities”, Proceedings of National Academy of Science, Vol. 79,
pp. 2554-2558.

[10] B. Shriver, ”Artificial Neural Systems”, Computer, March 1988, pp. 8-9.

141

 

[11] C. Chin, C. Maa, and M. Shanblatt, ”An Artificial Neural Network Algorithm
for Dynamic Programming”, International Journal of Neural Systems, Vol. 1,
No. 3, pp. 211-220.

[12] J. Hopfield, ”Neurons with Graded Response have Collective Computational
Properties like those of Two-State Neurons”, Proceedings of National Academy
of Sciences, Vol. 81, pp. 3088-3092.

[13] P. Bozovsky, ”Discrete Hopfield Model with Graded Response”, ICJNN Confer-
ence on Neural Networks, Vol. III, pp. 851-856.

[14] J. Hopfield and D. Tank, ””Neural” Computation of Decisions in Optimization
Problems”, Biological Cybernetics, Vol. 52, pp. 141-152.

[15] C. Keshavachandra and M. Shanblatt, ”Modeling Artificial Neural Networks Us-
ing VHDL”, Technical Report, MSU-ENGR-90-009, Michigan State University
Technical Report.

[16] R. Lipsett, C. Schaefer, and C. Ussery, VHDL: Hardware Description and

 

Desig_n, Kluwer Academic Publishers, 1989.

[17] D. Tank and J. Hopfield, ”Simple ”Neural” Optimization Networks: An A/ D
Converter, Signal Decision Circuit, and a Linear Programming Circuit”, IEEE
Transactions on Circuits and Systems, Vol. OAS-33, No. 5, pp. 533-541.

[18] User’s Manual for the Standard VHDL 1076 Support Environment, Intermet-
rics, 1988.

[19] W, Texas Instruments, 1984, pp. 3-224 and 3-227.

[20] The TTL Data Book, Volume 2, Texas Instruments, 1985, pp.3-4 and 3-5.

 

142

 

“IW]]]]]]]]]]]]]]“