M

:1thWill.lullﬂﬂ’ililelt

”NIH

I

 

N\
xalo LIBRARY I
MICI‘IC'V! State
University

I
W

 

This is to certify that the
dissertation entitled

SYNTHESIS OF PLANAR MECHANISMS
SIMULTANEOUSLY FOR BOTH
TYPE AND DIMENSIONALITY USING EVOLUTIONARY
SEARCH AND
CONVERTIBLE AGENTS

presented by

JOHN C. OLIVA

has been accepted towards fulﬁllment
of the requirements for the

Ph.D. degree in Mechanical EngineerinL

 

 

 

// ”AI/gawk

Ma]; Proidssor’ 5 Signature

y T /

 

Date

MSU is an Afﬁrmative Action/Equal Opportunity Employer

—.—.--.-.-.n-.—.-

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KzlProleocaPrelelRCIDateDuejndd

 

 

SYNTHESIS OF PLANAR MECHANISMS SIMULTANEOUSLY FOR BOTH
TYPE AND DIMENSIONALITY USING EVOLUTIONARY SEARCH AND
CONVERTIBLE AGENTS
By

John C. Oliva

A DISSERTATION
Smeitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Mechanical Engineering

2010

\TTT‘ ”“

ABSTRACT
SYNTHESIS OF PLANAR MECHANISMS SIMULTANEOUSLY FOR BOTH
TYPE AND DIMENSIONALITY USING EVOLUTIONARY SEARCH AND
CONVERTIBLE AGENTS
By
John C. Oliva
In the ﬁeld of mechanical engineering, synthesizing a mechanism to

perform an intended task is deceptively complex. In this dissertation, a novel
approach to automated mechanism synthesis is described which uses an
evolutionary search algorithm and a technique called “convertible agents” to
simultaneously ﬁnd the most appropriate mechanism type for a given problem,
while ﬁnding an optimum set of dimensions for that mechanism to realize a
speciﬁed behavior. The convertible agent technique has been developed in
response to the unique design Challenges encountered when synthesizing a
mechanism for both type and dimensionality. Several case studies are presented
which demonstrate the approach’s effectiveness over earlier solution strategies.
In these studies, six different planar single-degree-of-freedom mechanism types
are considered: a four-bar mechanism, Stephenson’s six-bar-mechanisms (types
I, II, and III), and Watt’s six-bar-mechanisms (types I and II). The method is
readily scalable to account for any number of different mechanism types and
complexities. The developed convertible agent approach is well suited for
evolutionary design applications outside of mechanism synthesis in which there
are a small number of distinct topological design possibilities each with

parametric variables to be optimized.

DEDICATION

This dissertation is dedicated to my wife Lisa. Without her constant support and

encouragement, this work would have never been completed.

iii

«9:; '~*" "‘

ACKNOWLEDGMENTS

I would like to thank Dr. Erik Goodman for his willingness to accommodate
my non-traditional degree trajectory throughout my tenure at Michigan State
University. When I was originally seeking a research advisor, he was unfeigned
by the prospect of working with a PhD candidate that intended to have a full-time
job outside of the university.

During my time at Michigan State University, I have held two professional
positions that allowed me both the ﬂexibility and the ﬁnancial support to pursue
my degree while gaining valuable career experience. This opportunity came ﬁrst
at Grand Valley State University's School of Engineering under the direction of
Paul Plotkowski and Jeff Ray (now with Southern Polytechnic University). Later,
a similar opportunity was provided by the National Superconducting Cyclotron
Laboratory thanks to the consideration of Thomas Glasmacher and Don Lawton.

I would like to thank all of the students that I had the privilege of teaching
at both Kettering University and Grand Valley State University. It was those
students that made me fall in love with engineering education and inspired me to
pursue my doctoral degree.

Finally, I would like to thank my parents, John and Eleanor, my sister
Jenni, and my Aunt Toni for their guidance and support throughout my education.
The initial direction that they provided from my early Childhood set my course for

this achievement.

iv

Table of Contents

List of Tables ...................................................................................................................... vi
List of Figures .................................................................................................................... ix
1.0 Introduction - A Historical Perspective ......................................................................... l
2.0 Mechanism Synthesis Overview ................................................................................... 5
3.0 Evolutionary Computing Overview ............................................................................ 10
3.1 Parallel Genetic Algorithms ................................................................................... 14
3.2 Multi-Agent Optimization Tool ............................................................................. 16
4.0 Mechanism Synthesis Using Parallel Search Agents .................................................. 18
4.1 Convertible Agents ................................................................................................. 22
4.2 Mechanism Evaluation ........................................................................................... 26
5.0 Case Studies ................................................................................................................ 36
5.1 Hinge Mechanism Case Study ............................................................................... 37
5.2 Walking Leg Mechanism Case Study .................................................................... 41
5.3 Challenge Mechanism Case Study ......................................................................... 43
5.4 Case Study Summary ............................................................................................. 45
6.0 Further Convertible Agent Studies ............................................................................. 47
6.1 Enhanced Convertible Agent Approach ................................................................. 47
6.2 “Known Solution” Case Study ................................................................................ 53
7.0 Conclusions ................................................................................................................. 60
Appendices ........................................................................................................................ 63
Appendix A - Mechanism Variable Maps ......................................................................... 64
Appendix B - Hinge Mechanism Case Study - Raw Data ................................................ 68
Appendix C - Leg Mechanism Case Study - Raw Data .................................................... 82
Appendix D - Challenge Mechanism Case Study - Raw Data .......................................... 96
Appendix E - “Known Solution” Mechanism Case Study - Raw Data ........................... 110
References ....................................................................................................................... 113

List of Tables

Table 1. Result Compilation for Furniture Hinge Mechanism Case Study. .................... 39
Table 2. Result Compilation for Leg Mechanism Case Study. ........................................ 42
Table 3. Result Compilation for Challenge Mechanism Case Study. .............................. 44
Table 4. Summary of Convertible Agent Average Objective Function Values for the
Three Case Studies (data repeated from earlier tables). ............................................ 45
Table 5. Result Compilation for the “Known Solution” Mechanism Case Study. .......... 56
Table 6. Hinge Mechanism - Independant Agents - Four-Bar Agent Results. ................ 69
Table 7. Hinge Mechanism - Independent Agents - Stephenson I Agent Results. .......... 70

Table 8. Hinge Mechanism - Independent Agents - Stephenson 11 Agent Results. ......... 71

Table 9. Hinge Mechanism - Independent Agents - Stephenson III Agent Results ......... 72

Table 10.
Table 11.
Table 12.
Table 13.
Table 14.
Table 15.
Table 16.
Table 17.
Table 18.
Table 19.
Table 20.

Table 21.

Hinge Mechanism - Independent Agents - Watt I Agent Results. ................... 73
Hinge Mechanism - Independent Agents - Watt 11 Agent Results ................... 74
Hinge Mechanism - Linked Agents - F our-Bar Agent Results. ....................... 75
Hinge Mechanism - Linked Agents - Stephenson I Agent Results .................. 76
Hinge Mechanism - Linked Agents - Stephenson 11 Agent Results. ............... 77
Hinge Mechanism - Linked Agents - Stephenson 111 Agent Results. .............. 78
Hinge Mechanism - Linked Agents - Watt I Agent Results. ........................... 79
Hinge Mechanism - Linked Agents - Watt 11 Agent Results. .......................... 80
Hinge Mechanism - Convertible Agent Results ............................................... 81
Leg Mechanism - Independent Agents - Four-Bar Agent Results. .................. 83
Leg Mechanism - Independent Agents - Stephenson I Agent Results. ............ 84
Leg Mechanism - Independent Agents - Stephenson 11 Agent Results ............ 85

vi

Table 22. Leg Mechanism - Independent Agents - Stephenson III Agent Results. ......... 86
Table 23. Leg Mechanism - Independent Agents - Watt I Agent Results. ...................... 87
Table 24. Leg Mechanism - Independent Agents - Watt 11 Agent Results. ...................... 88
Table 25. Leg Mechanism - Linked Agents — Four-Bar Agent Results. .......................... 89
Table 26. Leg Mechanism — Linked Agents - Stephenson 1 Agent Results. .................... 90
Table 27. Leg Mechanism - Linked Agents - Stephenson 11 Agent Results. ................... 91
Table 28. Leg Mechanism - Linked Agents - Stephenson 111 Agent Results ................... 92
Table 29. Leg Mechanism - Linked Agents - Watt I Agent Results. ............................... 93
Table 30. Leg Mechanism - Linked Agents - Watt II Agent Results ............................... 94
Table 31. Leg Mechanism - Convertible Agent Results. ................................................. 95
Table 32. Challenge Mechanism - Independent Agents - Four-Bar Agent Results. ........ 97
Table 33. Challenge Mechanism - Independent Agents - Stephenson I Agent Results. .. 98
Table 34. Challenge Mechanism - Independent Agents - Stephenson II Agent Results. 99
Table 35. Challenge Mechanism - Independent Agents - Stephenson III Agent Results.
................................................................................................................................. 100
Table 36. Challenge Mechanism - Independent Agents - Watt I Agent Results ............ 101
Table 37. Challenge Mechanism - Independent Agents - Watt 11 Agent Results. ......... 102
Table 38. Challenge Mechanism - Linked Agents - Four-Bar Agent Results. .............. 103
Table 39. Challenge Mechanism - Linked Agents - Stephenson I Agent Results. ........ 104
Table 40. Challenge Mechanism - Linked Agents - Stephenson 11 Agent Results. ....... 105
Table 41. Challenge Mechanism - Linked Agents - Stephenson 111 Agent Results ....... 106
Table 42. Challenge Mechanism - Linked Agents - Watt I Agent Results. ................... 107
Table 43. Challenge Mechanism - Linked Agents - Watt II Agent Results ................... 108
Table 44. Challenge Mechanism - Convertible Agent Results. ..................................... 109

vii

Table 45. “Known Solution” Mechanism - Original Convertible Agent (6) Results. 111

Table 46. “Known Solution” Mechanism - Enhanced Convertible Agent (24) Results.112

viii

Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.
Figure 9.
Figure 10.
Figure 11.
Figure 12.
Figure 13.
Figure 14.
Figure 15.
Figure 16.
Figure 17.
Figure 18.
Figure 19.
Figure 20.

Figure 21.

List of Figures
Typical Four-Bar Mechanism Conﬁguration. .................................................... 6
Example Mechanism Chromosome. ................................................................. 11
Mechanism Types Considered in This Study. .................................................. 18
Schematic of Six Independent, Parallel Mechanism Search Agents. ............... 19
Schematic of Six Linked Mechanism Search Agents. ...................................... 20
Illustration of an S3 Agent Having Been Converted to an 82 Agent. .............. 23
Mechanism Chromosome for Convertible Agents. .......................................... 28
Representative Watt I Mechanism. ................................................................... 30
First Step of Mechanism Construction. ............................................................ 32
Second Step of Mechanism Construction. ...................................................... 33
Third Step of Mechanism Construction. ......................................................... 34
Design Space and Target Locations for Furniture Hinge Mechanism ............ 38
Design Space and Target Locations for Leg Mechanism. .............................. 41
Design Space and Target Locations for Challenge Mechanism. .................... 43
Design Space and Target Locations for “Known Solution” Mechanism. ...... 55
Four-Bar Mechanism Variable Map. .............................................................. 65
Stephenson I Mechanism Variable Map. ........................................................ 65
Stephenson II Mechanism Variable Map ........................................................ 66
Stephenson III Mechanism Variable Map. ..................................................... 66
Watt I Mechanism Variable Map .................................................................... 67
Watt 11 Mechanism Variable Map. ................................................................. 67
ix

-"

1.0 Introduction - A Historical Perspective

To appreciate the concept of evolutionary mechanism synthesis, one must
ﬁrst understand the evolution of machine design itself. The story of man’s most
recent evolutionary steps cannot be separated from the evolution of his basic
understanding of the physical world around him.

Not long after man ﬁrst discovered the wheel, he attached some form of
pivoting axle to this wheel and the ﬁrst mechanism had been invented [1].
Archeologists and historians believe that the proverbial invention of the wheel
happened very early in modern man’s history, dating to around 3500 to 3000
3.0. By some deﬁnitions, it was an invention of this sort that ﬁrst distinguished
man from the primate world [2]. Harnessing the potential of what had been
discovered was something entirely different, though. It took thousands of years
for man to truly realize the potential of the mechanical curiosities he had
stumbled upon. Although the existence of mechanisms dates back to antiquity,
the formal study of their behavior and design is a rather recent practice.

For millennia after the invention of the wheel, most mechanical
innovations continued to be based on the wheel. The earliest chapters of
recorded history note widespread use of early wheeled machines like water
wheels and potter’s wheels. Beyond that, though, there was not much interest in
machines and mechanisms. During the time periods of the Greek and Roman
Empires, many mechanical contrivances were intended more as amusements
and curiosities rather than as means of doing productive work [2]. The one

enduring exception was in the tools of warfare.

An early ﬁeld of machine design that did captivate the interest of primitive
machine designers was the search for the design of a perpetual motion machine.
Given modern perspectives, the very notion of a machine that could run forever
without any outside sources of power or motion seems absurd. But dating back
to the ﬁfth century and continuing forward to the great minds of the likes of
Leonardo da VinCi (1452-1519), man sought this elusive self-propelled machinery
[1]. Of course, none of these experimenters were ever successful in their efforts,
and as a result, much of their time that could have advanced the study of
machines ended up being less productive.

Although the true scientiﬁc philosophers of early time periods did not want
to be bothered by the practical applications of science to the physical world, I
some advances were still made. Early machines and mechanisms were often
created by untrained inventors who sought to alleviate the burdens of their
everyday manual labor. These early machines were typically powered by natural
means, making use of windmills, animal power, or perhaps water wheels located
in nearby rivers [2]. It was not until the industrial revolution of the nineteenth
century that mechanisms as we know them today ﬁrst emerged and ﬂourished.
Two forces came together during this time period to open the doOr to a
resurgence of interest in all things mechanical. First, this time period saw the
initial widespread use of sources of rotational motion such as steam engines,
internal combustion engines, and eventually the electric motor, that were so
critical for reliably powering mechanisms. Furthermore, the increasing

mechanization of many agricultural, manufacturing, and other previously labor-

intensive processes made the study of mechanisms essential. Coupled with
these societal trends was the then budding profession of engineering, setting the
ground to establish the nineteenth and twentieth centuries as the glory days of
mechanisms. Many historians have even gone so far as to label the last 250
years of our civilization as “The Machine Age” [3].

Early investigations into the behavior of mechanisms approached the topic
from an analysis point of view. Engineers of the day studied existing machines
with the intent of being able to quantify their performance. This built upon the
scientiﬁc foundations laid before them by Galileo, Newton, and Lagrange to be
able to calculate the trajectories, velocities, and accelerations of machine
components they witnessed in operation before them.

The problem remained, though, not only to study an existing machine, but
also to be able to design a mechanism to perform an intended task without
having to employ the time-consuming method of trial and error so often used to
arrive at the earliest machine conﬁgurations. This process of mechanism
synthesis was a deceptively complex task that has continued to elude designers
through to modern times.

Throughout the last one hundred years, a plethora of schemes has been
proposed and implemented to synthesize a mechanism to perform an intended
task. Techniques have ranged from exhausting cataloging efforts whereby it was
intended to record all possible linkage conﬁgurations [4], to intricate graphical
procedures that could be followed to design a machine to meet certain input

parameters [5]. As engineering evolved into the applied scientiﬁc ﬁeld it is today,

analytical mathematical methods emerged [6]. Mechanism synthesis techniques
have continued to receive attention even into the computer age. Countless
computerized engineering optimization schemes have been adapted to solve the
mechanism synthesis problem.

Although such attention has been paid to mechanism synthesis
throughout the years, there still exists tremendous potential for improvement of
the synthesis tools available to modern practicing engineers. Many commercial
software packages are available to today’s engineer for the purposes of
analyzing mechanisms. But, just as their counterparts one hundred years earlier,
engineers are often "left to their own devices" when faced with the task of
synthesizing a mechanism. At present time, there are few commercially
available mechanism synthesis programs, and those that do exist have seen
rather limited usage. One might cite different reasons for these observations,
such as the lack of versatility of these programs, or a missing user friendliness
that profeSsional engineers have come to expect. Others have even gone so far
as to say that using a computer program to arrive at a design is counterintuitive
to an engineer's normal mode of operation [7]. Whatever the reasons may be,
the ﬁeld of mechanism synthesis continues to be one that could beneﬁt from the
development of better synthesis tools that could effectively design a mechanism

like those that have captivated man’s curiosity from the dawn of time.

‘4— -—

2.0 Mechanism Synthesis Overview

From a mechanical engineering point of view, the term “mechanism” refers
to a collection of mechanical components that are connected to each other via
some type of mechanical joints. The term “linkage” is often used interchangeably
with “mechanism”. By driving such a mechanism with a motor, engine, or
hydraulics, the mechanism is used to perform useful work. Although in reality
any value of load acting on a mechanism will cause the individual links to deform
from their unloaded shape, when designing and analyzing a mechanism, these
structures are typically considered to be completely rigid. With the rigid body
assumption, much of mechanism analysis becomes a practice in applied
geometry. The study of the Changing geometry of a mechanism as it passes
through its operational conﬁgurations is termed “kinematics".

Mechanisms can be classiﬁed in several ways. If all elements that make
up a mechanism and all of its resulting movements are conﬁned to a single plane
of motion, then the mechanism is considered a “planar" mechanism. Any
motions that take the links of a mechanism into a third dimension of action render
it a “spatial” mechanism. A mechanism can also be classiﬁed by the number of
links that it consists of. The simplest possible conﬁguration is a four-bar
mechanism. Note that it is customary to consider in this count the stationary
ground link that all of the moving links are attached to. Thus, to look at a four-bar
mechanism, one may observe that it has only three moving parts, but the fourth
bar is the immovable base or structure that the other three links are mounted to.

A schematic view of a generic four-bar mechanism conﬁguration is shown here

View-I. -0 3

as Figure 1. More complex mechanisms can be devised then by simply adding
to the link count, totaling ﬁve, six, or even more members within the mechanism.
With each additional component added to the linkage, it becomes more complex
to design and control, but the resulting motions can become that much more

complex as well.

D

Figure 1. Typical Four-Bar Mechanism Conﬁguration.

Designing a mechanism to meet a particular set of design constraints
involves two distinct activities. The ﬁrst is the mechanism synthesis component
of the process. Mechanism synthesis involves the speciﬁcation of the geometry
of the mechanism. Determining what type of mechanism to use (planar, spatial,
ﬁve-bar, etc.), what type of joints to use, the size of the links, and the
arrangement of the pivot points are all parameters deﬁned during the mechanism
synthesis stage of the design process. Once a candidate design has been
synthesized, then the designer moves on to the analysis stage whereby the
techniques of kinematics are applied to quantify such critical design

characteristics as the velocities and accelerations of points of interest on the

m 11.1 1" ..

machine, and simply the trajectories of the various points on the mechanism as it
passes through its range of motion.

The goals of mechanism design will of course vary depending on the
intended and use of the machine. Most authors, though, cite three distinct
classiﬁcations of mechanism synthesis. The ﬁrst is path generation, where the
goal is to have a point on the machine follow some speciﬁed trajectory. In the
second, motion generation, the position and orientation of one of the actual
member links is to be speciﬁed. The third classiﬁcation, function generation, is
where the mechanism is thought of as a conversion device, such that, for
example, the input force or angular movement is converted to a different output
force or movement [8].

Of the two tasks within mechanism design, the more complex portion is
the synthesis stage. As mentioned earlier, many tools are available to the
engineer when analyzing an existing mechanism. The geometric model of the
machine will provide “exact” determinations of the mechanism’s kinematic
properties. Commercially available software packages are also in wide use to
aid in this activity, regardless of how complex a given linkage may become.

Synthesizing a mechanism to meet intended speciﬁcations is a far more
complex activity. For synthesis, the mechanism designer is left without a well-
deﬁned algorithm. Although many methods of mechanism synthesis have been
proposed throughout the years, most offer only approximate results to varying
degrees of accuracy. Many of these techniques are tedious and laborious,

sometimes resulting in very poor designs. When this occurs, the entire process

VM‘ ’A‘ .

must be repeated either with different inputs throughout the process, or using an
entirely different technique.

Some of the earliest methods of mechanism synthesis were graphical in
nature. These techniques rely on drafting tools to physically sketch out the
synthesized mechanism through a series of geometric construction operations.
Graphical methods are readily available for three- and four-precision-point
requirements, where a point on the resulting mechanism will pass through the
speciﬁed number of points [6]. Although initially developed for hand-drawn
constructions, these techniques are now equally adaptable to computer—aided
design (CAD) tools. Variations on these same techniques exist for both motion
and function generation objectives. Analytical methods are also options for these
same purposes. Some of the purely math-based tools are directly based on their
graphical counterparts whereas others were developed independently.
Chebyshev’s, Bloch’s, and Freudenstein’s methods are all examples of analytical
mechanism synthesis schemes [5, 6].

Another method of mechanism synthesis attempted in the mid-twentieth
century was the development of coupler curve atlases. The concept here was
that if an exhaustive collection of mechanisms' behaviors could be cataloged,
then it would be a simple matter for the mechanism designer to ﬂip through the
pages of such an atlas and ﬁnd a mechanism that met the constraints speciﬁed.
The most enduring of these efforts is Hrones and Nelson’s 1951 Analysis of the

Four Bar Linkage in which some seven thousand linkages were documented [4].

The computer age ushered in a resurgence of interest in mechanism
synthesis research. Problems previously too complicated or with a vast number
of constraints were now within reach using computational methods. It seems that
practically every computer-based optimization algorithm has been applied to the
mechanism synthesis problem. Traditional gradient-based methods, pattern
matching, neural networks, and goal programming, are just a sampling of the
computer tools that have been used in the attempt to be able to ﬁnd the perfect

linkage for a given problem [9, 10, 11, 12].

3.0 Evolutionary Computing Overview

' Within the last couple of decades, another wave of interest has hit the
mechanism synthesis community with the introduction of evolutionary computing
techniques. Evolutionary computing is a ﬁeld of computer science that uses
various models of natural biological evolution as optimization strategies. This
class of techniques ﬁrst emerged in the late 1950’s [13]. But, as with preceding
advances in other disciplines, it would take many years for the new strategies to
be applied to mechanism synthesis.

The most basic form of evolutionary computing, as ﬁrst proposed by John
Holland, is a genetic algorithm (GA) [13]. The GA framework is a generalized
model for adaptive systems that has found applications in ﬁelds as diverse as
engineering, business, social science, and ironically, biological modeling. As the
GA optimization strategy strives to mimic biological evolution, the terminology
between the two paradigms also shares a lot of similarities. The fundamental
technique and the pertinent terminology will be illustrated here using a simple
mechanism application. Consider again the basic four-bar mechanism that
appears as Figure 1. The physical mechanism can be represented genetically as
a chromosome of genes. Programmatically, this may be an array of real
numbers. Here, the mechanism can be represented using a chromosome of
seven genes; two corresponding to each of the mechanism’s two pivot point
locations in X and Y Cartesian coordinates, and a length dimension for each link

of the mechanism. This chromosome is shown as Figure 2.

10

J .

 

Pivot A, X-Coordinate

 

Pivot A, Y-Coordinate

 

Pivot D, X-Coordinate

 

Pivot D, Y-Coordinate

 

Link Length AB

 

Link Length BC

 

Link Length CD

 

 

 

Figure 2. Example Mechanism Chromosome.

At the beginning of the evolutionary process, an initial population of
candidates is generated. A common initialization technique is to have the value
of each gene randomly generated by the evolution algorithm. (Ranges of
acceptable values for these parameters would be speciﬁed while deﬁning the
problem so that only meaningful candidates are created.) The next step is to
evaluate the ﬁtness of each member of the population. For the mechanism in
Figure 1, its ﬁtness may be measured by how well its motion matches an
intended motion. The techniques used to evaluate mechanism fitness for this
work will be detailed in section 4.2.

Three characteristic elements of a genetic algorithm are selection,
crossover, and mutation. Selection is the process by which two members of the
population are chosen to act as parents to generate offspring members for the
next generation. A multitude of different approaches exists for making this

selection, but most are probabilistic, giving those members with greater ﬁtnesses

11

 

a greater chance of being selected to act as a parent. This is in accordance with
the often-cited Darwinian theory of “survival of the ﬁttest".

Once two candidates are selected as parents, crossover is the process
through which the genetic deﬁnitions of the two parents are recombined such that
the resulting offspring design contains some genetic resemblance to each parent.
Again, any number of crossover methods may be used, but the most basic is that
termed “one-point crossover” in which a random point is chosen along the
chromosome. The offspring inherits the genes from one parent up to that point,
and then crosses over to the other parent’s chromosome for the remainder of its
genetic deﬁnition [14]. Thus, in the four-bar mechanism example here, an
offspring design may inherit its pivot point data from one parent, and then its link
length deﬁnitions from the other parent. The selection and crossover processes
then continue until an entire new population of designs has been generated.

Finally, in order to preserve genetic diversity, a small chance of mutation
is also introduced into the algorithm. As with selection and crossover, myriad
different mutation techniques have been suggested in the evolutionary computing
literature. But, the basic premise is that each gene of every design in the
population is given a small probability that it will be mutated. If the gene is
selected to be mutated, then it may be multiplied by a random multiplication
factor such that it experiences an increase or decrease in its original value [14].

With the newly deﬁned offspring population deﬁned, one may proceed to
the next generation in different manners. One technique is to blindly use the new

population, completely replacing the previous generation. Another method is to

12

compare each offspring to one of its parents, and the individual with the greatest
ﬁtness is carried forward to the next generation. "Elitism" is the strategy whereby
the best individual in the current generation is carried forth into the next
generation unchanged so as not to risk taking a step backwards in evolutionary
progress [14].

During the last ten to ﬁfteen years, many researchers have applied some
variant of a genetic algorithm to the mechanism synthesis problem. Similar to
previous introductions of new technology to the discipline, the mechanisms
synthesized with GA’s were taken back to their simplest forms for early studies,
and then slowly more complex mechanisms and objectives were strived for [15,
16, 17, 18, 19, 20, 21].

Thus far, several references have been made in regard to making the
mechanism synthesis task more challenging. This can be done in several
different ways such as by expanding the search to include more complex
mechanisms like six- or eight-bar linkages, or accOmmodating mechanisms with
non-planar behavior. However, another increase in difﬁculty comes with
synthesizing both the type and dimensionality of the mechanism simultaneously.
The GA mechanism synthesis studies already mentioned have all focused only
on optimizing the dimensionality of a mechanism for a given task; the type of
mechanism was chosen a priori [15, 16, 17, 18, 19, 20, 21]. In recent years, Liu
has enhanced this effort and documented several studies demonstrating the
simultaneous type and dimensional synthesis of planar mechanisms using GA’s

[22, 23].

13

In all of these cases, the researchers have used a fairly standard GA.
Although Holland originally conceived of GA’s as being deﬁned with strictly binary
data [13], and these mechanism studies have used real number representations,
beyond that, the mechanism synthesis studies using GA’s to date have not gone
far beyond Holland’s original framework. The research presented in this
dissertation expands upon (although completely independently) the direction that
Liu was taking in simultaneously synthesizing both the type and dimensionality of
planar mechanisms, primarily by employing a more sophisticated evolutionary
computing architecture and advanced optimization methods to obtain far greater

efﬁciency in the synthesis process.

3.1 Parallel Genetic Algorithms

The term “parallel genetic algorithm” (PGA) can take on different
meanings based on context. One interpretation can be in terms of how the GA is
functionally implemented when being executed. As described in the preceding
section, once an entire population of candidate designs has been deﬁned, the
ﬁtness of each member of the population must be evaluated. In many instances,
the computing time devoted to function evaluations far outweighs that consumed
by the overhead of the GA itself. One method of speeding computation time is to
run the function evaluations in parallel either on multiple processors, or physically
on different computers [14].

Another form of PGA is one in which the evolution itself is carried out in a
parallel fashion, which may or may not be on different processors. This

methodology is inspired by Danlvin's observations on the Galapagos Islands.

14

Because the islands are geographically isolated, many endemic species were
able to evolve independent of evolution taking place in other parts of the world. It
is thought that some of these species may have not survived had they been
competing for the same resources as more powerful species on other continents.
From an optimization point of view then, the concept is to have multiple separate
populations (“islands”) of designs evolving independently so that a very strong-
perforrning design found early in a run in one population does not immediately
overpower the entire computation domain, which may still ﬁnd other fruitful
designs, albeit in a slower manner. This method enables a GA to overcome the
tendency to prematurely converge to locally optimal rather than globally optimal
solutions. The islands may also be allowed to pass good designs to each other
(“migration”) on a periodic basis [14].

Lin et al. expanded upon this type of parallel genetic algorithm with their
model that they termed an “Injection Island Genetic Algorithm (iiGA)”. In this
approach, the optimization scheme utilizes multiple parallel populations of
candidate designs as just described. The added search efﬁciency comes from
the fact that each island can be searching using a different representation of the
problem. For example, suppose that the optimization problem at hand involves
the structural optimization of a mechanical component. As each design is
generated, it may be analyzed using a ﬁnite element model which is then solved,
and the resulting calculated mechanical strength is then used to determine its
ﬁtness. Using an iiGA, one population of designs could be using a rather coarse

mesh of elements to represent the geometry, while a second island of designs

15

uses a much ﬁner mesh. The coarse population will be able to evaluate designs
faster but with less accuracy than the reﬁned mesh representation. The iiGA
thus exploits the speed of the coarse mesh models by using that population to
quickly explore the design domain. As that population ﬁnds designs that perform
well, it periodically passes that design information to the population running in
parallel which then reﬁnes those designs and further optimizes the geometry with
a narrower search space and its ﬁner mesh of elements. One can see how this
model could be used to encompass any number of islands searching at varying

resolutions of a problem to quickly hone in on optimal solutions [24].

3.2 Multi-Agent Optimization Tool

Many researchers have developed parallel evolutionary search algorithms
allowing search within multiple populations and occasional interchange of
solutions among them. Because the emphasis of the mechanism synthesis
research presented here was not on the particular evolutionary operators to be
used, but rather on the organization of the search agents and of the
simultaneous search of the mechanism topology and dimensionality spaces, it
was decided to use an available commercial software tool rather than writing
another such search package. Red Cedar Technology offers a commercially
available general purpose optimization software called HEEDS (for Hierarchical
Evolutionary Engineering Design System). This software is the backbone of the
mechanism synthesisdescribed herein. The user provides the function
evaluation portion of the design analysis while HEEDS takes care of the core

evolutionary optimization. Furthermore, a user deﬁnes how each design will be

16

represented as it is passed between HEEDS and the function evaluation
program, as well as how the ﬁtness of each design will be reported back to
HEEDS. HEEDS provides the user with a range of optimization strategies from
traditional optimization schemes like quadratic programming to evolutionary-
based methods like genetic algorithms. HEEDS also offers Red Cedar
Technology’s own proprietary search method, termed “SHERPA”. This
technique uses a combination of different search algorithms and has
demonstrated its effectiveness on several parametric optimization problems [25].
All of the mechanism synthesis studies presented later in this paper employed
the SHERPA search strategy within HEEDS.

Similar to the iiGA model, HEEDS allows the user to deﬁne multiple
search “agents" to perform an optimization task. These agents are analogous to
the parallel islands in the iiGA model. Likewise, as in the iiGA approach, HEEDS
then provides the option of “linking” the deﬁned agents together so that they can
share design information on a periodic basis during the optimization run. Just as
with an iiGA, the agents can use different representations of the problem
deﬁnition, and the user merely needs to deﬁne how one representation “maps”

into the other [25].

17

4.0 Mechanism Synthesis Using Parallel Search Agents

Parallel evolutionary search agents were used within HEEDS in
conjunction with custom-written mechanism evaluation programs to synthesize
planar single-degree-of-freedom mechanisms. Six different mechanism types
were taken into consideration for a series of different synthesis case studies.
The six types included a four-bar mechanism, Stephenson’s six-bar—mechanisms
(types I, II, and Ill), and Watt’s six-bar—mechanisms (types I and II). These
mechanism types will be referred to as “48”, “S1 ”, “32”, “S3”, “W1 ”, and “W2”,

respectively. The six mechanism types are illustrated schematically in Figure 3.

 

Watt II

Figure 3. Mechanism Types Considered in This Study.

18

As will be described, the techniques used here are directly scalable to
account for any number of different mechanism types; these six were chosen in
this study merely to demonstrate the effectiveness of the techniques.

Three different approaches were used when synthesizing the
mechanisms. The ﬁrst was to establish six independent search agents; one for
each mechanism type. This approach was comparable to that taken by many of
the earlier researchers who used conventional GA’s to synthesize a single type
of mechanism as discussed in section 3.1. Although all six types were being
evolved simultaneously within HEEDS, the agents themselves had no interaction
with each other, so the results would not be any different than if six independent
runs were conducted, each searching for a single mechanism type. The six

independent search agents are illustrated in Figure 4.

 

 

 

 

 

 

4B SI 82 S3 W1 W2 .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 4. Schematic of Six Independent, Parallel Mechanism Search
Agents.

The second evolutionary approach was to take these same six agents, but
to link them together so that they could periodically exchange design information
regarding the best designs that each had found thus far during the evolutionary
process. This linking is shown in Figure 5. Notice that a link is created from

every agent to every other agent to enable them to exchange information directly.

19

 

Sl

 

 

S3

 

4B

 

 

 

 

 

 

 

/ \
\ /

1 ‘ W

 

 

Figure 5. Schematic of Six Linked Mechanism Search Agents.

Later in this dissertation, detailed results from the case studies
investigated will be presented to demonstrate the relative effectiveness of the
different evolutionary methods discussed here. However, it is instructive to
preemptively discuss the results of the ﬁrst two methods in general terms
because those results provide the motivation for the third evolutionary method
that was developed.

Three broad observations can be made when examining the mechanism
synthesis results obtained when using independent parallel agents and those
found using linked agents. First, the independent agents serve as a baseline
population with which to compare the subsequent methods. As mentioned,
evolving the mechanism types completely independently is essentially a
reproduction of the efforts performed by prior researchers using genetic
algorithms for their optimization. However, instead of a canonical GA, the more
efﬁcient SHERPA search algorithm in HEEDS was used. In the case studies

performed in this research, the independent agents always arrived at acceptable

20

solutions, demonstrating the functionality of the ﬁtness evaluation codes
developed here.

A second observation comes from the comparison of the results obtained
using the independent agents versus those from the linked agents. In general,
there was not a change in quality between the two result sets. Thus, the
mechanism agents do not appear to gain any meaningful design information
when exchanging data with each other. The author speculates that this is
analogous to phenomena witnessed in the natural world. For example, two
species may share many topological features (as do, for example, dogs and
cats), yet the two are not compatible enough to enable reproduction between the
species. In the case of mechanisms, referring to Figure 3, it can be seen that $2
and W1 mechanisms share a lot of topological features; they both have two
ground pivot points, they both have three binary links, and two ternary links. Yet,
the linked agent evolution method results indicate that typically the parameters
that yield a good performing SZ mechanism for a particular problem do not
translate into a W1 mechanism that performs well on that problem.

Finally, it can be observed that for all of the case studies, and for both of
the evolutionary methods, certain design problems tend to converge to the same
mechanism type consistently as the optimum solution. That is to say that certain
mechanism types emerge that are particularly well suited for speciﬁc applications
compared to other mechanism types. Conversely, this same mechanism type

may be the worst-performing type for a different application. This demonstrates

21

the need to consider a range of possible solutions for any given problem because

the top performing type is usually not obvious to the designer.

4.1 Convertible Agents

The synthesis approaches described thus far, although developed
independently and contemporaneously with the work by Liu, differ only insofar as
the performance of the SHERPA search algorithm differs from the traditional
genetic algorithm approach followed by Liu [22, 23]. In light of the observations
made in the preceding section, it was desirable to seek a multi-agent architecture
that would enable the agents to interact in such a way as to provide a more
effective search for the particular problem domain being investigated here.
Although developed within HEEDS, the methods developed below could also be
used within the framework of a conventional parallel genetic algorithm, and could
be reproduced independent of the HEEDS context.

It was clear that for any given problem, particular mechanism types were
never arriving at good designs compared to their better performing counterparts
because those types were ill-suited for the particular design objective. As a
result, the available computing resources were not being used as effectively as
they could be, since, inevitably, computing time was being consumed in every
run on mechanism types that were destined to be underperforrning from the
beginning.

The strategy for structuring the evolution that was developed to overcome
these shortcomings was termed “convertible agents”. In this scenario, the

evolution begins just as it did before, with six search agents running in parallel,

22

each looking for a speciﬁc type of mechanism to meet the common design
objectives. After a subset of evaluations has taken place, however, the best
results that each agent has found are compared. At this point, the agent that has
found the worst-performing mechanism to date is converted to become another
agent searching for the type of mechanism that has thus far found the best
performing mechanism type. This agent conversion event is illustrated below as
Figure 6. Even though all of the agents are always linked together in this
strategy, this ﬁgure emphasizes the added power that comes with the linking of

like agents in the search.

 

 

 

 

 

 

 

4B 81 82 ~ — 82 W1 W2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 6. Illustration of an 83 Agent Having Been Converted to an 82
Agent.

Thus, the computing resources that were being invested in a mechanism
that would likely be underperforming for the entire optimization run have now
been reallocated to searching the design space that has proven thus far to be the
most promising. Furthermore, throughout this process, all of the agents are
linked together and are intermittently exchanging design information with each
other. As noted eariier, dissimilar agents do not gain much from this exchange;
however, when two agents of the same type are linked, they progress with their
search faster than either agent can do alone.

This process of “converting” search agents continues as the search
progresses. Thus, after another subset of evaluations has been performed, the

new worst-performing mechanism agent is converted to be an agent of the type

23

that has performed the best. Eventually, as more agents of the same type
emerge, the linking of the like agents becomes more like that shown in Figure 5,
so that there are a number of agents all seeking mechanisms of the same type
that are all exchanging design information with each other.

HEEDS does not provide a means of changing the objective function of an
agent after an optimization run has begun. Therefore, the convertible agent
concept was implemented by the author by developing a program that runs in
parallel to the HEEDS internal optimization algorithm. This program essentially
monitors the progress of the evolution as it is happening. The monitoring
program maintains a list of which mechanism type each of the six agents is
currently evolving. Then, the mechanism evaluation code in any agent evaluates
each candidate design as any of the six mechanism types based on the type
deﬁnition it receives from the monitoring program. The monitoring program is
added to the HEEDS project as an additional analysis deﬁnition, just as the
analyses for the six mechanism agents are deﬁned. However, the monitoring
program does not have any inputs or outputs deﬁned within HEEDS. So,
HEEDS runs the monitoring program during every evolutionary cycle sequentially
after each of the mechanism agents, but all of its activities are hidden from the
HEEDS point of view.

The convertible agent approach has some similarities to a model
proposed recently by Gustafson that he has called “The Speciating Island Model”
[26]. In this work, as a new topology is discovered through an evolutionary

approach, its species is allowed to evolve independently so as not to be

24

overpowered by better performing species that have already been developed.
Similar to what has been witnessed in the mechanism analysis described herein,
the species are kept independent because they do not in general produce
meaningful results via reproduction with other species. The key difference
between the convertible agent concept and that of the Speciating island model is
knowledge of the species possibilities. For the mechanism synthesis problem,
the number of possible topologies is known a priori; there are only a distinct
number of combinations of links that create a single-degree-of-freedom
mechanism. This is in contrast to the Speciating island model which is in effect
being used to discover possible species. In this respect, the two models contrast
sharply.

The case studies described later in this paper show that for mechanism
synthesis problems that tended to arrive consistently at a particular mechanism
type using independent search agents, the convertible agent approach not only
tends to converge to this same type, but also for the same amount of computing
resources, a substantially better solution is found. Similarly, for problems that
had more than one of the six mechanism types that consistently performed well,
the convertible agents converge to those solutions proportionately as well.

It is instructive to compare the amount of search performed with the
convertible agent approach on each of the six topologies. For illustration, let us
assume one of the cases studied in which one of the topologies consistently
outperformed all others. For a run totaling 2,500 evaluations per agent (so

15,000 total evaluations), after each of the six 417-evaluation periods, an

25

additional agent would be converted to evaluating that topology, so the total
number of 417-evaluation periods spent on that topology is the sixth Fibonacci
number, 21. That equates to 8,757 evaluations in the agent using the best
topology (21 x 417). Let us contrast that with the effort it would take to do that
number of evaluations on each of the six topologies, which is 6 x 8,757 or
52,542. The savings in computational effort over that scenario is 37,542; the
convertible agent approach uses less than a third as many total evaluations to
ﬁnd a solution of the same quality. The convertible agent run also beneﬁts from
the island-parallel organization further increasing its search effectiveness.

Note also that the proposed convertible agent approach for mechanism
synthesis is directly scalable to consider any number of different mechanism
types. Six agents were used in this study to demonstrate the effectiveness of the
technique; however, one could envision how other agents seeking mechanisms
of different topologies could easily be added to the model. Any number of
possibilities, including non-revolute pin joints, more rigid links, or non-planar
behavior, could be included with the speciﬁcation of additional agents and their
associated evaluation tools.

4.2 Mechanism Evaluation

A subtle element that makes the convertible agent technique possible is
the way that the mechanisms themselves are represented. A novel approach
has been taken whereby every mechanism type uses an identical chromosome.
Even though the different mechanisms have different numbers of deﬁning

geometric parameters, a Chromosome of eighteen genes was Chosen such that it

26

can accommodate all six types, and certain genes go unused depending on the
speciﬁc mechanism being represented. Furthermore, the convertible agent
monitoring program maintains a list of what mechanism type each agent is
evolving. Then, a single mechanism evaluation code is used that can evaluate
the chromosome as any of the six mechanism types based on the type deﬁnition
it receives from the monitoring program. A schematic of the convertible
mechanism representation is depicted as Figure 7. Notice that the eighteen
possible design parameters included in this chromosome are all simple

geometric dimensions of the mechanisms.

27

 

Pivot 1, X-Coordinate

 

Pivot 1, Y-Coordinate

 

Pivot 2, X-Coordinate

 

Pivot 2, Y-Coordinate

 

Pivot 3, X-Coordinate

 

Pivot 3, Y—Coordinate

 

Link Length 1

 

Link Length 2

 

Link Length 3

 

Link Length 4

 

Link Length 5

 

Link Length 6

 

Link Length 7

 

Link Length 8

 

Link Length 9

 

Link Length 10

 

Link Length 11

 

Angle of Target Link

 

 

 

 

Figure 7. Mechanism Chromosome for Convertible Agents.

28

Regardless of the speciﬁc evolution method implemented, all of the
techniques described rely on an evaluation of each design's ﬁtness for guidance.
Any one of the commercially available, general purpose kinematic solvers could
be implemented to perform the ﬁtness calculation. However, in this study, since
only relatively simple, planar, single-degree—of-freedom mechanisms are
considered, a custom program was written to process just the mechanism types
considered, for speed and simplicity. In general, the algorithm cycles a
mechanism’s input link through its full range of motion, and calculates the
distance from the target link on the mechanism to each of the individual targets at
each step of the motion. Although the different mechanism types have deﬁning
geometric characteristics that make them each unique, the ﬁtness calculation for
the six types follows the same basic approach. For the sake of illustration, a
general Watt 1 mechanism will be used to demonstrate this ﬁtness evaluation.

Appendix A contains illustrations that map the generalized variables listed
in Figure 7 to the actual geometric interpretation Of those variables for each
mechanism type. Consider the W1 mechanism depicted in Figure 8. As shown
in the appendix, the genetic information maps to this mechanism type as shown

via equations 1 through 16.

29

 

Figure 8. Representative Watt I Mechanism.

Ax = Pivot 1, X-Coordinate
Ay = Pivot 1, Y-Coordinate
Dx = Pivot 2, X-Coordinate
Dy = Pivot 2, Y-Coordinate
AB = Link Length 1
BC = Link Length 2
BF = Link Length 3
CF = Link Length 4
CD = Link Length 5
CH = Link Length 6
DH = Link Length 7
FG = Link Length 8

GH = Link Length 9

30

(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)

(13)

Ex = Link Length 10 (14)

Ey = Link Length 11 (15)

(p = Angle of Target Link (16)

The following development will demonstrate how the mechanism
conﬁguration of Figure 8 is arrived at. However, mechanism kinematics are not
easily visualized using solely a static ﬁgure. Because of this, it is helpful to
describe the motion of this mechanism prior to delving into the geometric
construction of it. In the ﬁgure, link AB can rotate a full 360° about the pivot point
at A. This is the mechanism's input or driven link. Links BCF, CDH, EFG, and
GH are connected to link AB through a series of revolute joints. Hence, as A8
rotates through its range of motion, CDH tends to oscillate back and forth about
pivot point D, while the other links move accordingly. Recall that this is a single-
degree-of-freedom mechanism as are all of the linkages considered in this study.
Thus, for any speciﬁed input angle of link AB, the rest of the mechanism's
conﬁguration is uniquely deﬁned. (It should be reemphasized that link EFG is
taken here as a single rigid body whereas the other points in the illustration are
revolute joints.)
When the evaluation code receives the mechanism deﬁnition from

l-lEEDS, it can directly plot points A and D. Furthermore, the evaluation code
speciﬁes an angle for the input link. This angle 9, along with the link length AB

enables the calculation of the coordinates of point B via equations 17 and 18.
Bx = Ax + AB- cos(e) (17)

By =Ay +AB-sin(6) (18)

31

At this point, the mechanism construction appears as Figure 9.

 

 

Figure 9. First Step of Mechanism Construction.

Making use of the Law of Cosines, and link lengths BC, CD, and 80, the
coordinates of point C can be calculated. First, an intermediate step is required
to calculate the length of the third side of the triangle that will be used in this
operation, in this case 30. This is a "virtual length" in the sense that the
dimension is actual, but there is no physical mechanism link that corresponds to
this direct connection. The geometric constructions developed via equations 19

through 23 are illustrated in Figure 10.

 

 

BD = J(Dx 49,)2 + (Dy 49),)2 (19)
2 2 2
a=cos_1{BC +31) —CD I (20)
I 2-BC-BD J
D —B
.—1 y y
= —— 21
I. ( .. ] .,
Cx=Bx+BC-cos(a+ﬂ) (22)

32

Cy = By + BC- sin(a+ ,6) (23)

 

Figure 10. Second Step of Mechanism Construction.

The same approach taken to locate C with respect to points B and D can
be utilized to construct the two ternary links of the W1 mechanism as well. Thus,
point F can be positioned with respect to points 8 and C, and point H can be
located with respect to points C and D. (Note however that in these cases,
because these are ternary links, the intermediate step comparable to calculating
the virtual length 80 is not needed, because all of the link lengths for the ternary
links are speciﬁed in the mechanism deﬁnition.) Once these calculations have

been performed, the mechanism construction appears as shown in Figure 11.

33

 

Figure 11. Third Step of Mechanism Construction.

The remaining pivot point of the mechanism, point G, can be located with
one more iteration of the trigonometric approach taken before, and again
calculating a new virtual length, this time length FH.

The remainder of the geometric development is rather straightforward.
The location of target link E is deﬁned via one dimension measured along link
F6, and a second dimension measured perpendicular to link FG (see equations
14 and 15). Finally, the angle (p speciﬁes the angle of the target link with respect
to link FG. Thus, the complete position of the mechanism has been determined
as was shown in Figure 8.

In the motion generation (or “rigid body guidance") mechanism synthesis
problems considered in this study, the objective is to have the target link E pass

through a series of prescribed positions speciﬁed to complete a given task [8].

34

Thus, a pair of X,Y point coordinates is also speciﬁed for each target position.
The program developed for this study can account for any N targets. The length
of target link E is taken to be equal to the target lengths themselves.

To arrive at the results presented later in this work, a step size of 1° for
the input angle 9 was used for completeness, although a larger step size could
have been used with a negligible loss of accuracy. Thus, at any given angular

step 9, the distance d from link E to target ican be deﬁned simply as:

 

di,6=\/(Ex‘ix)2+(Ey‘iy)2 (24)
This is actually done for the points deﬁning each end of the target and
target link respectively, and the total distance is taken as the sum of the two.
Throughout the range of motion, the distances are monitored, and the minimum
distance at which the target link approaches each target is recorded.

d; = min (dug) (25)
9=0°-—)359°

The mechanism’s total objective function then is taken as the sum of these

minimum distances for each of the N targets.

ObjectiveFunction = Edi (26)
i =l—>N

Additional code is included to allow for mechanisms that do not have a full
range of motion yet still produce meaningful results, and to eliminate
mechanisms that have only very limited ranges of motion by assigning arbitrarily

high objective function values.

35

YEW.“ “it“ E11

5.0 Case Studies

To demonstrate the relative effectiveness of the three different evolution
methods as applied to mechanism synthesis, a series of case studies was
investigated. These case studies were inspired by mechanism synthesis projects
carried out by earlier researchers and documented in the engineering literature.
It should be emphasized that it was not the intent of this effort to ﬁnd “better”
solutions than those originally presented in each work. Instead, these cases
were chosen because they represented real-world design tasks taken from a
variety of different ﬁelds, and they were all cases that posed some challenges to
the original investigators. In all of the cases though, the original work involved
design considerations that went beyond the scope of the current research. Thus,
only the kinematic motion generating properties of the cases were used as the
basis for the studies presented here. In some cases, however, the convertible
agent approach does demonstrate its power in the quality of solutions found, and
the fact that it proceeds with the push of a button as compared to the often
elaborate methods originally employed by the initial investigators.

Certain parameters were the same for all three case studies. In each
case, the problem at hand was solved using each of the three different evolution
methods: independent agents, linked agents, and convertible agents. Each
strategy was performed 25 times to gain conﬁdence in the consistency of the
results produced. (Since the evolution process involves a multitude of random
numbers, there is always the possibility that very good or very bad results could

be found as an anomaly during any particular run.) HEEDS was conﬁgured to

36

*.

evolve the mechanisms using 2500 calls to the objective function under the
SHERPA method. In the case of the converging agents, the monitoring program
was set to convert an agent every 417 iterations (one sixth of 2500) so that at the
conclusion of the synthesis, there was the possibility of having all of the agents
converge to a single type.

5.1 Hinge Mechanism Case Study

The ﬁrst case study comes from the ﬁeld of furniture / cabinet design. A
quick glance at any ofﬁce or home furnishings with doors shows that a lot of
effort has gone into the design of the hinges. The objective of the designers is to
have a hinge that provides a smooth range of motion while taking up as little
space as possible.

Chen presented a formal engineering approach to this design objective
whereby a mechanism was synthesized to act as a furniture hinge. The
synthesis goal was to arrive at a hinge that would initially have the door move
perpendicular to the door opening and then swing wide open to reveal the
contents behind the door. Furthermore, it was desired to have hinge
components that required as little mounting space as possible, and that would
allow them to be surface mounted to both the door and the main support
structure [27]. (Many modern hinges require elaborate mounting features to be
cut into the cabinet itself to allow proper functionality.)

During the original synthesis process described for the hinge design study,
a series of target positions was not explicitly deﬁned as is the approach in the

current research method. Thus, a series of ﬁve target positions were selected for

37

this study to replicate the motion as described by Chen. Other details such as
allowable link lengths, permissible pivot locations, and sizes in general were
chosen here to be consistent with Chen's work, although such speciﬁcs were not
provided in the original publication. The geometry of this design space is shown

in Figure 12.

 

 

 

 

 

Y (Inches)

 

 

X (Inches)

 

 

 

Figure 12. Design Space and Target Locations for Furniture Hinge
Mechanism.

In Figure 12, the origin is taken as the contact point between the cabinet
door and the cabinet structure. The cabinet would extend up along the +Y axis,
and the door would extend back along the -X axis, in its closed position. The ﬁve
lines plotted in the ﬁgure are the ﬁve target positions for the hinge through its
range of motion. They are connected with a dotted line to indicate the expected

path between them. Notice how the two targets closest to the cabinet are exactly

38

parallel with each other, yielding the opening motion as described earlier. Also
note that the ﬁrst target is inset from the door when it is in its closed position.
This is to meet the criteria that the hinge can be surface mounted. Finally, the '
box in the ﬁgure extending from the origin to the point (-0.5, 4.0) represents the
allowable range that was speciﬁed for the location of the ﬁxed ground points.

Appendix B contains a complete reproduction of all of the evolution results
found for this first case study. A summary of these results appears here as
Table 1. The ﬁrst three columns show the average objective function results
obtained when using independent agents, linked agents, and convertible agents,
respectively. The next two columns compare the effectiveness of the three
approaches using the independent agents as a baseline for comparison. Notice
the sign convention adopted for this comparison. A mechanism that passes
through the targets more exactly is indicated by a lower objective function value.
However, when comparing the average objective function values in this paper,
better results are indicated with a positive percentage comparison. The ﬁnal
column of Table 1 indicates the number of times the convertible agents

converged to each mechanism type during the 25 runs.

Table 1. Result Compilation for Furniture Hinge Mechanism Case Study.

 

 

 

 

 

 

 

 

Mech. Ind. Linked Conv. Ind. vs. Ind. vs. Conv. Agent
Type Agents Agents Aggits Linked Conv. Solutions

48 1.40 1.71 1.21 -22.1% 13.6% 2

81 2.58 1.66 0.99 35.7% 61.6% 8

82 1.81 1.81 0.61 0.0% 66.3% 9

83 2.54 2.10 0.99 17.3% 61.0% 3

W1 3.97 3.03 - 23.7% - 0

W2 2.35 2.06 1.15 12.3% 51.1% 3

 

 

 

 

 

 

 

 

39

‘27----22 -. .

With the presentation of the remaining two case studies, it will be shown
that this ﬁrst case study is the only one investigated for which signiﬁcant gains in
performance were achieved using linked agents over independent agents. As
shown in Table 1, four of the mechanism types saw respectable gains, while only
the $2 mechanism type remained constant, and the 4B mechanism type was
worse. The convertible agents found much better results for all of the
mechanism types with the exception of W1. The W1 type performed very poorly
as a furniture hinge mechanism using all of the synthesis methods attempted - so
poorly, in fact, that the convertible agents never converged to a solution that
used the W1 topology, hence its entry for the convertible agents is left blank in
Table 1, and it has a zero entry in the right most column.

It is also interesting to note once again that the W1 mechanism type was
found to be the worst type for this particular synthesis case by all three
evolutionary methods. This is notable because in the original work, Chen
concluded that the W1 type was the ”most feasible” for the hinge design [27].
However, Chen approached the problem from a different perspective, using the
Creative Mechanism Design Method to ﬁrst select the mechanism type, and then
optimizing the dimensions of that type. Chen did take into consideration design
aspects that went beyond the kinematic behavior sought here, but those
considerations were made after the mechanism type had already been chosen.
Here, by synthesizing the type and dimensionality of the mechanism

simultaneously, a very different conclusion is drawn.

4O

5.2 Walking Leg Mechanism Case Study

The second case study is taken from the robotics literature. Shieh et al.
describe an effort in which they synthesized a mechanism to act as a leg for a
walking robot. The desired path of the mechanism's foot is to follow a delta-
shaped trajectory so that it can climb up stairs [28]. As with the ﬁrst case study,
speciﬁc details about the objectives and limits for variables needed in the current
synthesis approach were not originally speciﬁed, so reasonable values were

speciﬁed here for the permissible lengths and bounding regions.

 

 

P---
b---

- 15

I ............ .0

Y (Inches)

.....
9".
'0
'0
'0
o
O
o
o
o
"
'
'
'
I
l
O
I
'
I
I
’I
I

 

§ 0
‘9‘- "O
Q ......
..........
-- --
.................
................

 

-1o -8 -6 4 -2 o 2 4 6 8 10
x (Inches)

 

 

 

Figure 13. Design Space and Target Locations for Leg Mechanism.

Figure 13 shows the targets selected to trace out the delta-shaped foot
path. Notice that in this case, the robot's foot could be effective if it were a sharp

point, thus for this work, an arbitrarily short target length was chosen, leaving a

41

series of target positions that appear to be points instead of lines. The dotted
box at the top of Figure 13 represents the region within which the mechanism
pivots can be located, which would correspond to where the leg would be
mounted to the robot. From a kinematic perspective, this complicates the
synthesis task by having the mechanism's ﬁxed ground located remotely with
respect to the desired path of action.

When synthesized using HEEDS and the three different evolutionary
methods, the results shown in Table 2 were found. A complete listing of the raw
"data for this case study can be found in Appendix C. In this case, only the W1
mechanism type yielded considerable improvements when linked agents were
used instead of independent agents. However, once again, the convertible
agents produced much better mechanisms than either of the other two
evolutionary approaches, given comparable computing resources.

As witnessed in the ﬁrst case study, the search agents for this particular
synthesis study found one mechanism type to be consistently ill-suited for the
task. Here it was the $2 mechanism type for which the independent and linked
agents found the poorest results, and consequently then, the convertible agents

never found a solution of this type.

Table 2. Result Compilation for Leg Mechanism Case Study.

 

 

 

 

 

 

 

 

Mech. Ind. Linked Conv. Ind. vs. Ind. vs. Conv. Agent
Type Agents Agents Agents Linked Conv. Solutions

4B 8.68 8.41 4.48 3.1% 48.4% 8

S1 11.62 12.06 3.08 -3.8% 73.5% 4

82 24.09 27.19 - -12.9% - 0

S3 8.67 7.89 4.49 9.0% 48.2% 7

W1 11.52 8.94 6.57 22.4% 43.0% 3

W2 12.17 12.44 3.29 -2.2% 73.0% 3

 

 

 

 

 

 

 

42

 

5.3 Challenge Mechanism Case Study

The ﬁnal case study considered was originally proposed for the sake of

acting as a challenging case for mechanism synthesis approaches. Professor

Michael McCarthy of the University of California - Irvine challenged software

vendors to devise a mechanism synthesis tool that could design a mechanism

that would pass through eleven speciﬁed targets along an irregularly shaped path

[29]. A re-creation of this design space is shown in Figure 14.

 

Y (Inches)

 

 

 

 

1

<5:

 

 

 

 

.6 -
X (Inches)

Figure 14. Design Space and Target Locations for Challenge Mechanism.

Although the targets were distinctly speciﬁed in the original challenge, the

remaining constraints (pivot location range, allowable link lengths, etc.) were less

speciﬁc, and were thus chosen here by the author. Compared to the earlier two

43

 

case studies, the complexity of this present study comes from the sheer number
of targets, and the path that they lie along.

As with the other case studies, HEEDS was used to synthesize
mechanisms to pass through the eleven speciﬁed targets using the three
different evolutionary methods. The complete data set from this case study
appears as Appendix D. The average results obtained from these efforts are

recorded in Table 3.

Table 3. Result Compilation for Challenge Mechanism Case Study.

 

 

 

 

 

 

 

 

Mech. Ind. Linked Conv. Ind. vs. Ind. vs. Conv. Agent
Type Agents Agents Agents Linked Conv. Solutions
4B 9.33 8.91 5.26 4.5% 43.6% 9
S] 8.91 10.51 5.23 -18.0% 41.3% 2
82 9.13 11.39 7.34 -24.8% 19.6% 5
S3 9.36 10.25 5.88 -9.5% 37.2% 4
W1 10.98 10.83 - 1.4% - 0
W2 10.62 10.80 5.56 -1.7% 47.6% 5

 

 

 

 

 

 

 

Here, more-so than with either of the preceding case studies, the
convertible agent approach demonstrates its superior effectiveness compared to
the other evolutionary methods. That is to say, for the same amount of
computing resources (i.e. an equal number of function evaluations), the
' convertible agents consistently arrived at solutions that better met the speciﬁed
targets. Once again, linking the agents in this case caused the solutions to have
worse average objective function values than when the agents were allowed to
evolve independently, for every type except the 4B and W1 mechanism types,

which only achieved negligible gains with the linked agents.

44

 

5.4 Case Study Summary

As was hypothesized, the convertible agent evolutionary method
developed here for the synthesis of mechanisms has demonstrated through the
preceding case studies to be a robust method that consistently arrives at better
solutions than the other evolutionary methods investigated. Further insight
comes from a comparison of the results arrived at using the convertible agent
method for each of the three case studies. For convenience, these result sets

are summarized in Table 4.

Table 4. Summary of Convertible Agent Average Objective Function Values
for the Three Case Studies (data repeated from earlier tables).

 

 

 

 

 

 

 

Mech. Type Case Study 1 Case Study 2 Case Study 3
4B 1.21 4.48 5.26
S1 0.99 3.08 5.23
82 0.61 - 7.34
S3 0.99 4.49 5.88
W1 ' - 6.57 -
W2 1.15 3.29 5.56

 

 

 

 

 

 

The three case studies were intentionally presented in an order in this
dissertation indicative of their relative synthesis complexity. The relative
magnitudes of the average objective function values for each case illustrate
these differences. For example, in the ﬁrst case study, the furniture hinge, there
were only ﬁve targets arranged in a relatively simple path. Thus, the evolved
mechanisms very easily match those targets, and the average objective function
values are consequently rather low. Compare this to the eleven targets arranged
in an irregular path for the third case study, the Challenge mechanism, where the

best mechanisms tended to have comparatively higher objective function values.

45

Even if all of the values in Table 4 are normalized on the basis of objective
function value per number of targets, the results for the third case study are still
three to four times larger than for the ﬁrst case study, owing to the complexity of
mammapmn

Recognize also the variety of results obtained among the three cases in
terms of the mechanism types found to be best and worst. For example, in the
ﬁrst case study, the 82 mechanism was found to be the top-perfonning
mechanism type, on average. However, in the second case study, this same
mechanism type performed so poorly, on average, that the convertible agents
never arrived at a solution containing the 82 type. And, although there are such
differences among the results sets, there are also some notable similarities. For
example, the W1 type performed pooriy for all three of the case studies
considered in this research. In the ﬁrst and third cases, the convertible agents
never arrived at a W1 solution. In the second case study, even though some W1
solutions were found, on average, they were still the worst solutions arrived at for
that study. The author speculates that given enough characteristically different
synthesis cases, situations would emerge where each mechanism type is found
to be the best solution for the task, while other cases would be found where that
type is the worst solution. All of these Characteristics serve to further
demonstrate the beneﬁts of being able to consider automatically a number of

different mechanism topologies for any given synthesis task.

46

6.0 Further Convertible Agent Studies

The preceding two chaptershave sought to demonstrate the motivation
leading up to the convertible agent approach, the speciﬁcs of that methodology,
and the performance of the approach as applied to actual mechanism synthesis
problems. As with any system, after studying its behavior, one can propose
variations to further enhance its performance. This Chapter seeks to discuss
such efforts. First, several enhancements to the evolution approach are
proposed. Then, using this enhanced convertible agent strategy, the results from
a case-study synthesis problem for which there is a known solution are
presented to illustrate that elements of the original strategy have been improved.
6.1 Enhanced Convertible Agent Approach

A key characteristic of the convertible agent method is its ability to
reallocate underutilized computing resources being wasted on underperforming
solution types to those types that are demonstrating the most promising
convergence performance. However, in the approach as demonstrated thus far,
this reallocation is done decidedly unintelligently. That is to say that as the Six
convertible agents progress through a given optimization run, the worst
performing type gets converted to an agent type like that which is performing the
best. But, this ”best" and “worst" comparison is done blindly in regards to the
relative magnitudes of the agent's objective function values. So, at the point in
time of comparison, if all six of the agents are performing relatively equally, the
”best" and “worst" could differ by a ﬁnite yet trivial amount. Because the

convertible agent approach as illustrated thus far does not take this in to

47

consideration, the "worst" agent will still get converted (and thus eliminated) from
the population.

In a similar vein, the timing of the conversion event in the convertible
agent approach discussed thus far has been chosen rather arbitrarily. As
described in the case studies of chapter 5, the conversion interval was set to be
one sixth of the total number of evaluations performed during the full optimization
run. This was selected simply because there were six agents corresponding to
the six mechanism types considered, and it was desired to have the possibility of
having all of the agents converge to a single type at the conclusion of any
particular solution. Again, one can argue that this conversion decision is being
made based upon very little information. It could be that a particular agent may
have a relatively high objective function value compared to the other agents, but
it is making greater strides in improving itself compared to its peers. If that is the
case, it would be beneﬁcial to provide that agent with additional time to converge
to an optimized solution rather than eliminating it solely based on its objective
function value. In recognition of this phenomenon, a more rational approach
would be to decide agent conversion not on an arbitrary number of performed
evaluations, but rather on a measure of an agent's historical performance such
that agents that have stagnated become more likely to get converted.

To address the deﬁciencies just described, several enhancements were
made to the convertible agent approach to improve the criteria on when agent
conversions take place, and which agents get converted. First, to take into

consideration the relative values of all of the agents’ best objective function

48

values, at every iteration of the enhanced convertible agent method, the objective
function for each agent is recorded. With these values, the average objective
function value of the agents, and the standard deviation of this data set is
calculated. Agent convergence is then determined on an "as needed“ basis
based on the relative range of objective function values. Convergence can occur
if the “best” agent has an objective function that is a speciﬁed number of
standard deviations less than the average of the population, while the “worst"
agent's objective function value is that number of standard deviations greater
than the average of the population. In this sense, now the "best“ and “worst" are
not deemed to be so by trivial amounts, but because they are true outliers of the
overall population.

A series of experimental runs was performed to determine what measure
of standard deviation disparity would lead to productive agent convergence. If
the measure is too high, such that the “best" and "worst" agents need to be a
large number of standard deviations away from the average, then the agents
never get converted, and the search progresses as though the agents were
merely running in parallel. If the threshold is too low, then the agents are
constantly getting converted, and the search progression becomes undirected
and does not converge effectively. Through this trial and error process, it was
found that a value of 0.75 standard deviations yielded a search with fruitful agent
conversions.

Another element added to this enhanced convertible agent method is a

monitoring of the evolution progression of the different agents. Even with the

49

value comparisons as just described, one can imagine that in a given iteration of
the method, one particular agent may get converted. During the next round of
comparisons, this agent will have had only a mere fraction of the function
evaluations to shape its design that the other agents have had, so it is likely to
still have the worst objective function value of the population. If the convergence
criteria were limited to just objective function values, then this same agent would
get converted again and again throughout the search. To avoid this, the
enhanced convertible agent method also takes into consideration the iteration-
based gradient of each agent's objective function value. In the scenario just
proposed, even though the newly converted agent has the highest (i.e., worst)
objective function value, in general, at the beginning of any evolutionary process,
that agent will also likely be making the greatest strides in improving its
performance. Thus, during each iteration of the enhanced convertible agent
method, the objective function gradient with respect to number of iterations is
calculated for each agent based on its previous 100 reported objective function
values. Added to the convergence criteria already described then is a constraint
that the agent getting converted must have either the lowest convergence
gradient of the population (i.e., it is making the least progress) or its gradient has
to be zero, which could be the case in the event that all of the agents have
stagnated.

To summarize the convergence criteria in the enhanced convertible agent
method, consider the pseudo code below appearing as equations 27 through 29.

An agent gets converted if the following three criteria are simultaneously met:

50

ObjFuncmax > ObjFuncavg + 0.75 - 0' (27)
ObjFuncmin < ObjFuncavg — 0.75 - 0' (28)
VObjFuncmax < 0.00001 OR 170bjFuncmax = min(V0bjFunc) (29)

In these equations, ObjFuncmax, ObjFuncmm, and ObjFuncavg are the
maximum, minimum, and average objective function values of the set of all
agents in the population, respectively. The standard deviation of the set of

agent objective function values is denoted as o. The gradient of the agent with
the highest objective function value is shown as V ObjFuncmax and the minimum

agent gradient of the set of all agents is min(V ObjFunc). Note that this minimum

agent gradient is determined regardless of that particular agent's objective

function value, whereas V ObjFuncmax is determined speciﬁcally as the gradient
of the agent with the worst objective function value.

Notice that in equation 29, instead of determining a stagnated agent as
having a gradient of zero, a small yet ﬁnite objective function value is used in
practice so as to avoid numerical rounding issues. Furthermore, it should also be
mentioned that in the actual implementation of the enhanced convertible agent
method, no agent conversions are allowed during the initial start up period of a
particular run, or for a set number of iterations after an agent conversion occurs.
These periods were set at 1667 iterations, and 100 iterations, respectively. This
allows the agent behavior to settle a bit after the disruption that the agent
conversion causes. Again, these values were chosen based on experimentation

with different threshold values.

51

0.x- -- t I v

‘F'.

A ﬁnal enhancement to the convertible agent approach is in regards to the
number of agents employed. Much like some of the other parameters described
in this section, the number of agents was chosen in the original convertible agent
method to be six simply because that was the number of mechanism types being
considered. The error in this selection is due to the probabilistic nature of any
evolutionary approach. As illustrated in the case study results of chapter 5,
although any particular synthesis problem tends to have one or two mechanism
types that appear to be particularly well suited for that intended motion based on
the frequency of ﬁnding solutions of those types, the fact remains that during any
given run a particularly good or particularly bad design for any of the types may
emerge that can steer the results in a different direction. The logical way to avoid
this inevitable pitfall is to increase the population size such that anomalous
outliers tend to have less of an effect on the overall progress of the domain
search. For example, given the results of Case Study 1 shown in Table 1, one
can see that the Stephenson Il mechanism was the most often found solution by
the convertible agents. For the 25 solution runs performed, 9 times (or 36% of
the trials), the convertible agents arrived at a Stephenson ll type. If one
subscribes to the hypothesis that the Stephenson ll type is the best suited for this
particular mechanism synthesis problem, this data also indicates that more than
half of the time (64% of the trials) the convertible agents converged to a less than
ideal solution type. In light of this, by expanding the agent population to 24
agents such that there are initially 4 agents searching for each of the 6

mechanism types, then the probabilities indicate that any particular run is more

52

likely to arrive consistently at the optimum type rather than when there were only
6 agents. The subsequent section of this chapter exhibits results from a new
case study that utilized the enhanced convertible agent conversion strategy as
well as this increased agent population size.

6.2 "Known Solution" Case Study

In the previous chapter of this dissertation, the results of three different
case studies were presented to demonstrate the effectiveness of the convertible
agent evolution method for mechanism synthesis. In each of these studies, the
convertible agents arrived at effective solutions for complex synthesis problems
inspired by real-world design studies taken from the mechanical engineering
literature. Although "good" solutions were found to each problem, one cannot
conclude that a true optimum solution was found without undertaking an
exhaustive search of the design space at hand. For the case studies presented
thus far, a mechanism with a perfect objective function of zero was never found.
Instead, solutions very close to the objective were always arrived at. But, at the
same time, without undertaking the aforementioned exhaustive search, one
cannot even say with certainty that there is an ideal solution to be found within
the space being searched.

To overcome this ambiguity in the performance evaluation of the
convertible agent method, a fourth case study has been considered. This
problem was fashioned in the inverse sense to the preceding case studies, such
that the problem begins with a solution generated perfectly by a known

mechanism. In other words, the problem deﬁnition began with a known

53

mechanism design that was selected at random. The motion of this mechanism
was then characterized, and a series of target positions that it passed through
were recorded. With those targets now acting as the mechanism synthesis
objective, the problem was presented to the convertible agents just as before.
However, in this case, regardless of the quality of solutions found, it is known that
a perfect solution does indeed exist.

One additional element was added to this case study problem deﬁnition to
bolster its credibility. As mentioned in section 5.4, a trend recognized in
comparing the results from the three case studies was that the Watt 1 mechanism
type did not perform well for any of the case studies considered. The hypothesis
proposed was that any given synthesis problem has a particular mechanism type
that is well suited for the required motion, and then conversely, other types will
be particularly ill-suited for the problem at hand. Seemingly, just by chance, the
three case studies chosen all happened to share some characteristics that led
the Watt 1 type to be unsuited for those applications. To put this theory to the test,
this fourth case study began with a Watt I mechanism type that deﬁned the
perfect target locations. If the convertible agents were to arrive at consistent
solutions utilizing the Watt l conﬁguration, not only does it demonstrate the
effectiveness of the solution strategy, but it also strengthens the arguments
regarding the concepts that different mechanism types are better for certain
problems than others.

The desired target positions for the fourth case study are shown here as

Figure 15. Notice that there are a total of 33 targets. This is a large number of

54

targets compared to the preceding studies that had a maximum of 11 target

positions.

 

-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18
12

 

Y (Inches)
.z\
'7
l
l
I
l
’7‘!
, r
0)

 

1
—(
.1
J
.4
—(
M

 

 

 

 

X (Inches)

 

 

 

Figure 15. Design Space and Target Locations for "Known Solution"
Mechanism.

This case study was attempted with two different evolutionary strategies;
the 6-agent architecture like used for the ﬁrst three case studies, and the 24-
agent approach described in the foregoing section that employs the more
sophisticated convergence algorithm. Each of these two methods was run 25
times and the results compiled. Note also that for the sake of comparison, the
total number of function evaluations was kept equal between the two methods, at
a total of 60,000 evaluations. (That is, in the 6-agent case, each of the agents

completed 10,000 evaluations, and in the 24-agent case, each of the agents

55

performed 2,500 evaluations.) The compiled results from this effort are shown

here as Table 5. The complete raw data from these runs are contained in

 

 

 

 

 

 

 

 

 

 

Appendix E.

Table 5. Result Compilation for the "Known Solution" Mechanism Case

Study.

Mech. Average Objective Function Values Number of Convergences
Type 6 Agents 24 Aggits 6 vs. 24 6 Agents 24 Agents

48 - 28.69 - 0 3

S1 39.87 - - 2 0

S2 37.26 32.88 11.8% 2 4

S3 27.26 28.92 -6.1% 14 7

W1 25.10 27.35 -9.0% 6 8

W2 36.10 26.40 26.9% 1 3

 

 

 

 

 

 

 

The results of this latest case study exhibit a combination of different
features. First, in terms of the solutions found, recall that the known perfect
solution for this demonstration was a Watt I type. The original 6-agent
architecture did indeed ﬁnd on average solutions of the Watt I type to have the
best performance. With the enhanced convertible agent methodology and 24
search agents, the Watt I type ranks second, lagging only behind the Watt II type.
In consideration of the fact that the earlier case studies never converged to Watt l
mechanisms, the argument that these strategies can converge to the best type
for a given synthesis problem has been bolstered.

In comparing for each mechanism type the average objective function
values found by the two different methods (the fourth column of Table 5), one
sees inconclusive results. For a third of the mechanism types (Stephenson II
and Watt II), the enhanced convertible agent methodology did arrive at better

results than the original implementation as proposed. For another third of the

56

types (Stephenson Ill and Watt I), the enhanced method actually arrived at worse
results than the original method (albeit, by a smaller margin than the previous
agents were improved by). Finally, for the remaining two mechanism types (Four
Bar and Stephenson I), one method or the other did not converge to a
mechanism of one of those types, and thus there is no basis for comparison.

The last two columns of Table 5 illustrate the frequency with which each
mechanism type was converged to. It is evident that for this particular case
study, the code often converged to both Stephenson Ill and Watt 1 types. With
the original convertible agent method and just 6 agents, the search converged to
a Stephenson Ill type a majority of the time (56% of the trials). The supposition
was that by adding additional agents to the search effort along with the enhanced
convergence strategy, the agents would have more of a tendency to consistently
arrive at the known optimum solution. Strictly speaking, the 24-agent case did
demonstrate that the most often found solution was a Watt I type as expected,
but the Stephenson lll mechanism type still performed consistently well.

It should also be pointed out that neither method ever arrived at the true
optimum solution. The 6-agent implementation did arrive at a single Watt I
solution with an objective function value of 8.89, and upon inspection of the
variable values, there is some resemblance between that design and the known
perfect design. Thus, one can conclude that these techniques produce usable
designs that for practical applications Come close to a designer's intended
motion, but one cannot claim that the techniques are able to ﬁnd a globally

optimum mechanism design for a speciﬁc problem.

57

Recall that the number of function evaluations for each trial was selected
such that all runs used an equal amount of computing resources so as to
maintain an even basis for comparison. However, it should also be emphasized
that in a majority of the runs (approximately 85%), the top performing design was
found at the end of the maximum number of allowed iterations. There is no
reason to believe then that if given additional iterations, the agents would not
have continued to reﬁne and improve the optimum solutions even further.

As the discussion leading up to this case study explained, there are many
contributing factors to the behavior of the enhanced convertible agent
methodology. An overriding dynamic is the problem dependence of the method.
It seems that the case study here, which as noted, was speciﬁed somewhat
randomly, just happens to lend itself to a perfect solution of a Watt I type, but
apparently simpler to ﬁnd (although less than ideal) solutions of the Stephenson
Ill type. So, as with the previous case studies, without performing an exhaustive
search of the design space, one cannot Characterize the number of solutions that
truly exist. It may be that the one known perfect Watt I solution for this synthesis
problem is somewhat at a point of instability such that none of its Close relatives
have a performance that resembles its own. Conversely, the Stephenson lll type
might have a lot of variations that come close to matching the intended behavior,
but none exist that meet it perfectly.

All other control parameters of the enhanced convertible agent method
could be further optimized to produce better results. The values chosen for

standard deviation weighting, initial convergence interval, intermediate

58

 

convergence intervals, et cetera, could all be studied at length to ﬁnd what
combination yields the most productive search. However, one may also ﬁnd that
this, too, is problem dependent.

In the end, it appears that the slight performance gains achieved are not
worth the added computational expense of the enhanced convertible agent
method along with the expanded population of search agents. But, this is not to
say that further reﬁnement of the enhanced method would not lead to greater
efﬁciency gains. It is likely that some hybrid between the original convertible
agent method and the enhanced approach may be the preferred strategy. One
could envision using the enhanced convergence strategy with the simpler 6-
agent implementation to improve convergence behavior while not incurring
additional overhead. Another element that could beneﬁt further investigation is
the degree to which the search agents are linked together to exchange
information so that the network promotes collaboration, but not immediate

homogenization of the population.

59

7.0 Conclusions

In 1995, renowned mechanisms expert Arthur Erdman wrote a
commentary on the current status of computer-aided mechanism design as part
of the American Society of Mechanical Engineering's 50th anniversary
commemoration. After presenting an overview of what was then considered to
be the state-of-the-art in terms of mechanism design software, he concluded that
"The great majority of software addresses kinematic and dynamic analysis
needs...Software for kinematic synthesis is less available and less mature."
Furthermore, he proposed that "In the future one can hope for either a general-
purpose mechanical system dimensional synthesis package or a series of
separate but connected software tools that are optimized for each class of
mechanical systems“ [7]. Today, a decade and a half after Erdman made these
assessments, one can argue that the state of the computer-aided mechanism
design world of the practitioner has not advanced far from where it was in 1995.
Although research in mechanism synthesis techniques has remained vibrant, and
strides have been made in all manner of computing capabilities, mechanism
synthesis tools have remained elusive to the practicing engineer. Newly
developed tools seemingly remain conﬁned within the research facilities of
academia. Perhaps it is a consequence of the techniques being out of reach for
the professional mechanism designer, whether that be for reasons of complexity,
computing overhead, or mathematical abstraction. Maybe it is simply a matter of
the mechanism synthesis community not posing a great enough need to garner

the attention of major software developers.

6O

The point remains that synthesizing mechanisms is an engineering design
problem that has both captivated and frustrated generations of engineers. The
work presented in this dissertation demonstrates that one can now build upon the
efforts of earlier mechanism synthesis researchers to devise fully automated
synthesis strategies that simultaneously optimize both the type and
dimensionality of a mechanism.

The studies presented in this dissertation have illustrated two key
guidelines that mechanism designers can beneﬁt from regardless of speciﬁc
solution strategy. First and foremost, the results presented demonstrate the
value in considering a variety of solution topologies; often times the best
mechanism type is not obvious. In that same vein, this research has challenged
the commonly held belief in the mechanism synthesis world that more complex
mechanism types will always yield better performance than simpler types. In
some of the case study results presented, it was shown that the best type also
happened to be the simplest. Thus, the convertible agents approach each
problem without any predisposed biases; much like a human designer should.

Additionally, the convertible agent approach developed herein is readily
scalable to consider any number of different mechanism types that go beyond
those considered in this initial study. Mechanisms with different joint types,
increased numbers of links, or non-planar behavior can be directly added with
the speciﬁcation of the respective search agents. With each additional type
considered, the designer will gain conﬁdence in knowing that they are using the

mechanism type best suited for a particular application.

61

 

From an evolutionary computing point of view, the convertible agent
approach lends itself to be applied to problems outside of the mechanism
synthesis task for which it was developed. The technique is well suited for
evolutionary design applications in which there are a small number of distinct
topological possibilities, each with parametric variables to be optimized. This is
in contrast with genetic programming, which has often been applied to synthesis
of open-ended system topologies and associated parameters. The convertible
agent approach is dramatically more efﬁcient, as the space of topologies (or
mechanism types, in this case) that must be searched is small enough to be
searched, at least preliminarily, in an exhaustive manner. As with any
evolutionary approach, the convertible agents and enhanced convertible agents
have many control variables (conversion rates, comparison thresholds,
stagnation gradients, and the like) that need to be optimized or “tuned” to
improve performance for a general class of problems.

The convertible agent evolutionary approach developed in this research
has been shown to be a more robust and efﬁcient evolutionary synthesis strategy
for mechanisms than any of the methods that have come before it. It represents
a methodology that could be put into service for real-world mechanism design
applications with relative simplicity, providing ﬂexible capabilities to deliver a
variety of useful results. To this end, the synthesis community can beneﬁt from
the further advancement of such tools to ﬁll a void that has been apparent for

generations.

62

Appendices

63

Appendix A - Mechanism Variable Maps

64

 

Figure 16. Four-Bar Mechanism Variable Map.

Len 8

/
Iken 11

    

Figure 17. Stephenson I Mechanism Variable Map.

65

 

Figure 19. Stephenson Ill Mechanism Variable Map.

66

Put 1

 

Figure 20. Watt I Mechanism Variable Map.

 

Figure 21. Watt II Mechanism Variable Map.

67

Appendix B - Hinge Mechanism Case Study - Raw Data

68

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.. 00.. 000 00.0 000 00.0 00.0 00.0 00.0 .00- 00.0 00
00.. 00. 00.0 000 00.0 0... 00.0 00.9 00.0 000. 00.. 00
.0. 00.. 00.0 00.0 000 00.0 00.0 00.0 00.0 00.0 000 00
.0. 00.. 00.0 00.0 00.0 00.0 00.0 0.0- 00.0 .0? 0... 00
.0. 00.. 000 0... 00.. 00.0 00.0 00.0 000 0.0- 00. .0
.0. 00.. .00 00.0 .00 000 000 00.9 000 .00. 00.0 00
.0. 00.. 00.0 .00 000 00. 0.0 020 00.. 00.0. 00. 0.
.0. 00.. 00.0 0... 00.. 000 00.0 0.? 00.0 000. 00.. 0.
.0. 00.. .00 00.0 00.0 00.0 00.0 00.0 00.0 .00. 00.0 0.
.0. 00.. .0. 00.0 00.0 00.. 0.0 .00- 00.0 .00. 000 0.
.0. 00.. .00 00.0 00.0 00.0 00.0 00.0 00.0 .0? 00.0 0.
.0. 00.. 00.0 00.0 00.0 00.0 00.0 .00. 00.0 0.0. 000 0.
00.. 00.. 00.0 00.0 .00 .0. 00.0 00.0- 00.0 00.0 000 0.
.0. 00.. 00.0 000 000 00.0 000 .00. 00.0 00.0 00. 0.
.0. 00.. 00.0 000 00.0 000 000 00.0 000 00.0. 00.0 ..
.0. 00.. 00.0 00.0 00.0 00.. 00.0 000- 000 .0? 00.. 0.
.0. 00.. 00.0 .0. 00.. 00.0 0.. 00.9 00.0 .0? 00.. 0
00.. 00.. 00.0 00.0 000 00.0 00.0 00.0 00.0 000. 00.. 0
00.. 00.. 00.0 00.0 00.. 00.0 0.0 .00. 00.. .0? 00.0 0
00. 00.. 00. .0. 00.0 000 00.0 000- 00.0 .00. 00.0 0
.0. 00.. 00.0 .00 0.0 00.0 00.0 ..0. 000 00.0 00. 0
00.. 00. 00.0 00.0 00.0 0.0 0.0 00.9 00.0 00.9 00. 0
.0. 00. 00.0 .00 00.0 00.0 000 00.9 00.0 00.? 00.. 0
00.. 00.. 00.0 .00 000 000 00.0 00.9 00.0 .00. 00.0 0
00.. 00.. 00.0 00.0 00.0 00.. 00.0 00.0 00. 00.0- 00.. .
0.050 0W. 0"; 0w; 0W. 0%. “mm. a.“ 9.... .0“... 2.00.00 :3.

 

 

.3500”. 0:090 Lame-son. .. 3:004. Eugene—E. .. 8&5...er 09:... d 030...

69

 

00.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

50.. 00._. 60.0 006 50.0 006 006 00.0 05.0 00.0 006 66.0 00.0- 000 N00. N60 0N
50.0 00.0 00.N 00.N 50.0 00.N 50.0 00... 0N.N 06.N 0.20 00.N 00.0 00.0- 0N.0 00.0. 00.0 6N
00.0 00... 00.0 :N 06.N 00.. 06.0 00.0 06.. 00.N F5.N N00 60.0 56.0. N00 NN.0. :20 0N
50... 00.0 5N.0 N00 00.0 :N 00.0 00.0 50.0 006 SN 50.0 0N.0 00.0 N0... 00.0 056 NN
50.0 00. P 05.0 006 00.0 00.0 N00 50.0 00.0 00.N 00.0 006 06.0 0N0. 60.0 06.0- 006 N
50. w- 00... N0.0 50.0 60.0 mrN 50.0 0N.0 06.0 05.0 00.0 006 00.0 06.0- 0—20 00.0. N50 0N
50.0 00. —. 00.. 00.N 05. w 60.N 00.N 65.N 60.0 65.N 00.0 05. r 05. r 50.0. N0.0 00.0. N0.N 00
50.. 00. F 0N.N 50.0 06.0 50.N 60.0 00.0 00.N 60.0 50... 00.N 00.N 0.20- 000 00.0. EN 0..
60.0 00._. 05.0 66.0 N00 05.0 006 06.0 00.0 00.0 00.N N5.0 00.0 06.0- 000 00.0 50.N 5..
00.0 00. F 00.0 0N.N 006 00.0 006 05.0 5N 50.0 006 00.N 00.0 00.0. 00.0 00.0 00.0 00
00... 00.0 00.. 60.0 50.0 06.N 006 006 00.N 00.0 00.0 00.N 60.0 00.0 00.0 N60- 00.0 0..
50.. 00. 0 00.0 006 50.0 00.N 60... 60.N 00.0 05.N 5N.0 50.N 00.0 00.0- 00.0 00.0. 00... 6.
00.0 00. 0 N00 00.. 00.0 N0.N 50.0 00.N N00 05.0 00.0 00.0 60.0 00.0 N_..N 00.0- 06.0 00
00,—. 00. _. N00 EN 006 60... 5N0 00.N 50.. 50.0 00.0 006 00.0 00.0 00.0 No.0. 06.0 N.
50... 00.0 00.0 60.N 00.0 00.0 00.N 00.N 006 50.N No.0 00.0 N00 00.0- 000 0N0. 0N.N 00
06.0 00. _. 0N. F 00.0 00.N 0N. r 50.N 05.0 N09 N00 0 w.N 00.0 0N0 00.0 00.0 00.0- 00. w 0.
50.. 00.0 00.0 00.0 N00 06.0 NN.N 0N.0 60.0 006 00.N 00.0 00.0 00.0- N0.0 N00- 006 0
60.0 00. F 05.0 65.N 06.0 50.0 00.0 or... 00.0 00. F 00.0 50. w N...0 00.0- 00.0 No.0- 66.0 0
50... 00. r 60.0 06.0 00.0 05.0 006 00.0 50.0 06.. 006 00.N 66._. N60. N00 00.0- 05.0 5
00.0 00.0 N0.0 6N0 00.0 N0.N 00.0 mwN 0N.N N00 60.N 00.0 60.0 00.0. N00 5N0- N00 0
0N. .. 00. _. 00.0 00. F 0N0 60... 006 00.0 00.N 00.0 006 05.0 00.0 00.0 00. _. 00.0- 00... 0
50.0 00. w 05.N 60.N 006 05. w 006 0N.N 00.0 0 r .N 00.0 006 00.0 0N0. 00.0 06.0. 05.0 6
50.0 00. r 006 60.0 006 00.0 006 006 006 00.0 00.0 006 00.0 00.0- 00.0 00.0- 0 w .0 0
60.0 00.0 00.0 0N0 00.0 00.N 50.0 50.0 00.0 50... 00.0 F5.N 00.0 $.0- 000 0N.0- 00.0 N
60.0 00. _. 00.0 60.N 60... 00.0 00.0 06.0 05. F 00.N 006 00... 00.0 06.0- 00.0 66.0- 00.0 r
202 CFO—”d COO—w— :“J C“.— CWJ CWJ C“.— C".— and CM:— C":- ”C“ “XE.“ ”a“ 0X6“ SSE—Do =3“

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.3303. 0cou< . cameo—380 - £ceu< 0:00:33:— - 80:23: 35: .5 030...

70

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.0 50.0 00.0 0N.0 00.N 05.N N00 00.N 60.0 06.N 00.N 00.0 00.0 0N.0. 00.0 00.0- 05.0 0N
00. 0- 00. 0 50. 0 00.0 N0. 0 00.0 00.0 0N.0 05.N 60.0 No.0 05.0 N00 N0.0- 00.0 00.0- 50. 0 6N
50. 0 00. 0 00.0 00.N 0N.N N0.N 00.0 00.6 00.0 5N.0 00.0 00.6 00.0 00.0 00. 0 5N.0. 00.0 0N
06.0. 00. 0 00.0 00.N 50.0 60. 0 NN.N 0 0.N 50.0 00.0 06. 0 06.0 0N.0 0 0.0. 00.0 00.0- 00... NN
50. 0- 00. 0 00.6 00.0 00.0 00.0 65.N 06.0 00.N 00.N 00.0 00.6 00.0 06.0. N0.0 00.0- 00.N 0N
60. 0 00. 0 00. 0 00.0 N0. 0 0 0 .N 00.0 06.N 06.N NN.N 00. 0 00. 0 N5.0 00.0- 05.0 00.0- 50.0 0N
06. 0 50.0 00. 0 00.0 5N.0 05.N 50. 0 06. 0 60.0 0N.0 50.N 00.0 N0.N N60. 00.N 00.0. 00.0 00
00.0 50.0 00.0 N0.0 00.0 00.0 05.0 60.0 00.0 06.0 05.0 N5.0 00.0 00.0. 05.0 00.0- 00.0 00
0N.0 00.0 60.0 0N.0 50.0 0N.N 0N.N N5.0 65.N 50.0 00.N 05.N 60.N No.0. 60.N 06.0- 50.N 50
00.0- 00. 0 06.0 N0. 0 50. 0 No.0 06.0 0 0.0 00.0 06.0 5N. 0 50.0 60.0 00.0. 00.0 50.0. 00. 0 00
0N.0 00.N N0. 0 00.N 05.0 50.N 00.N 60.N 0N. 0 00.0 0N.0 00.0 00.0 60.0- 600 Nd. 00.0 00
66.0 00.0 00.0 60.N N5.0 00.6 00.0 05.0 50.0 00.0 00.N 50.0 06.0 6N.0. 00.N 00.0. 05.N 60
00.0 50.0 00.0 00.0 00.0 00.N 60.0 00.0 00.0 00.0 00.0 50.N 00.N 00.0. 00.N 00.0. 05.N 00
00.0 00.N 60. 0 0N.N 00.6 5N.0 00.N 6N. 0 N5.0 6N. 0 05.0 00.0 00.0 66.0- 00.0 0N.0. 06.0 N0
50. 0- 00. 0 N0.N 50.0 00.6 00.N 00.6 00.6 00.6 00.6 00.6 00. 0 0N.0 00.0 00.0 00.0. 0N.0 00
0N.0 00.0 00.0 05.N 00.0 05.0 6N.0 06.0 00.N N0.N 00.0 50.0 66.0 0N.0. 00.0 N6.0. 60.N 00
00.0 00.N N0.N 00.N 00.N 00.N 00.0 0N.0 05.0 60.0 00.0 00.N 60.0 00.0. 06.0 0N.0. 00.0 0
00.0 00. 0 00.N 00.0 60.N 06. 0 00.N 60.N 6N. 0 5N.0 00.0 05.N 60.0 66.0- 60.0 06.0- 00.0 0
06. 0 00. 0 05.N 50. 0 N0. 0 50. 0 05.0 00.0 50.N 5 0. 0 00. 0 00.N 00. 0 60.0. 00.0 06.0. No.0 5
0 0. 0 00.N 05. 0 00.N 05.0 05. 0 06.0 65.N 05.N 0 0.N 00.0 00.N 00.0 00.0- 00. 0 00.0- 00.0 0
00.0- 00.0 60.0 00.0 00.N 60.0 00.N 00.N 00.N 00.0 00.0 00.0 05.0 00.0 60.0 00.0. N0.N 0
00.0- 00.0 60.0 00.0 00.0 00.0 50.0 05.0 60.0 05.N 0N.0 60.N N00 00.0 66.0 00.0- 05.0 6
50. 0- 00. 0 05.N 0N.0 50.0 00.0 00.0 N00 00. 0 00.0 N0.N 00.N 60.0 00.0.. 0N.N 5N.0- 00.N 0
N0. 0 00.6 00.0 00. 0 00.6 00.0 0 0.0 00.N 00.N N0.N 00. 0 06.N 00.0 6N.0- 00.0 06.0. 00.N N
50.0 00.0 N5.0 00.0 5N.0 05.0 05.0 65.N 60.N 00.0 66.0 00.0 00.0 00.0 00.0 00.0- N0.0 0
.0... ..... .0. .0 ....._ .0 .0 ..... .0 .0 .0 0.... .0. 0... .0. 0......

 

 

.3300”. acoo< = concegnoum - 3:004. aceucaouE - EmEusoos. O05... .0 030...

71

III-I. . III-II. 1.12.0...

._

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.0. 00.. 00.0 00.. 000 000 00.0 00.0 00.0 00.0 00.. 00.0 00.0 00.0- 00.0 00.0 00.. 00
.0. 00.. .0. 00.. 00.0 00.0 000 00.. 00.0 .00 000 0.? 00.0 00.0 00.0 00.0- 000 00
00. 00.. 00.0 .00 00.. .00 .0. 00.. 00.. 0.0 0.0 .0? 00.. 00.0 00.0 00? 0.0 00
00. 00.. 00.0 000 00.0 .00 0.0 0.0 0.0 00.0 00.0 000. 00.0 00.0- 00.. 000- 00.0 00
00. 00.. 00. 00.0 .00 00.0 00.0 .00 000 000 00.0 00? 00. 00.0. 0... 00.0- 0... .0
.0. 00.. 00.0 00.. 000 00.0 .0. 000 00.0 00.0 00.. 00.0- 00.0 00.0. 00.0 000. .0. 00
.0. 00.. 00.0 00.0 00.0 000 000 00.. 000 00.0 00.0 000- 00.0 000. 00.0 00.0. 00. 0.
00.. 00.. 00.0 000 0.0 0.0 00.0 000 00.0 .0. 00.0 00.0 00.0 00.0. 00.0 0.0- 00. 0.
.0. 00. 00.0 000 .00 .00 00.0 00.. .00 000 000 00.0- 0.0 000- 00.0 00? 00.0 t
.0. 00.. 00.0 000 00.0 00.0 000 .00 00.0 .00 0... 00.0- 0... 00.0. 00.. 00.0- .0. 0.
00.. 00.. 00.0 00. 00.0 0.0 00.. 00.. 000 000 00.0 00? 0.0 00? 000 00? .0. 0.
00.. 00.. 000 00.0 .0. 00.0 ...0 00.0 .00 .00 00.0 00? 0.0 00.0- 00.0 00? 00.. 0.
0... 00.. 00. .00 .00 000 .0. 000 00.0 00.. 00.0 000. 00.0 00.0- 00.0 0.0- 000 0.
00.. 00.0 000 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0- 000 00.0 00.0 00? 000 0.
.0. 00.. 00. 00.0 .0. 00.0 000 00.0 .00 000 00.0 00.0. 00.0 000- 00.0 00.0- 000 ..
.0. 00.. 0.. 00.. 0... 00.0 000 00.0 0.0 00.. 00.0 .00- 000 000- 00.. 00.? 000 0.
.0. 00.. 00.. 000 00.. .00 000 .00 00.0 00.0 000 00? 00.. 00.0- 00.0 00.0- .00 0
00.. 00.. 000 00. 000 000 000 00.0 00.0 00.0 00.. .00. 000 ...0. 000 00.0- 00.0 0
.0. 00.. 00.0 000 .00 00.0 .00 .00 00.0 00.0 000 0.? 00.0 .0? 00.0 00.0 000 0
00.. 000 ..0 00.. 00.. 00.. 000 00.. 00.0 00.0 00.0 000. 000 .0? 00.. 00? .00 0
00. 00.. 00.. 000 .0. 000 000 000 0.0 .00 00.0 00? 0.0 00.0- 00.0 00? 00.. 0
00.. 00.. 0.0 00.0 00.0 00.0 .00 000 00.0 000 00.0 8.0 00.0 0.? 00.0 00.0- 00.. 0
.0. 00.0 000 00. 00.0 000 0.0 00.. 000 00.0 000 00.0- 00.0 000- 00.0 00.0. .00 0
00.. 00. 00.0 00.. 00.0 00.. .00 .00 00.0 00.0 00.0 00.0- 00.. 0.0- 00.0 00.0. 000 0
.00 00.. 000 00.0 00.0 000 00.. .0. 000 00.. 00.0 0.? 00.0 00.0. 00.0 00.0. 00.0 .
0.92 a”; 0W. 0%.. cw... emu cu; 0m. 0m. an. IR. w...“ 9m ““1 PM a... 0.5030 03.

 

 

.00.-50¢ acou< _= cameo—330 . 3:090 «co—0:300:— - 80.50.35. 005: .0 030...

72

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

smé 00... 0N.—. 06.N 006 00.N 06.N 00.0 00.N mm... 00;. or... 60.0 00.0. NP... 00.0. 06.0 mm
96.0 00% are 00.0 006 N0.N 00.0 00.0 00.0 00.0 05.0 00.0 00.? 00.0 00.0 0 F0. 56 6N
so; 00... 09.0 00.0 Sum 00.« 0N.N 05.0 00.0 00.0 5.0 60.0 66d 6N.0. 66.? 06.0. 00.0 mm
50.0 00. F 00.0 00.N 00.0 00.N 3.0 «0.0 50.0 50.0 5.0 006 N5.0 00.0. 60.0 00.0. 5.0 mm
mm... 00. P 00.0 0N.N 50.0 «N? 00.0 00.N 00.N 00.0 00.0 05.? 00.N 00.0 00.0 06.0- 5.0 _.N
wmé 00. r 00.0 50.0 00.0 60.0 00.? *6; mm.» 00.0 006 wvé 00.0 3.0- «0.0 00.0- 0N.0 0N
N09 00._. 50.0 00.N sud 00.N 00.0 50;. 00.N 00.0 00.? 00.0 6N.0 00.0 00.0 0N.0- NN.N 09
um... 00.? 006 00.0 006 60.0 006 006 30.0 0N.0 00.0 0rd N50 5.0. 00.0 00.0. 00.0 0..
bmé 004 60.0 50.0 00.0 0N.N 00.0 60.0 066 No.0 0N.0 66.0 006 00.0. 60.0 N00. 006 t
60.0 00... 00.0 60... 50.0 006 N5.0 N5.0 00.0 60.0 66.0 00.0 9.0 «0.0. 3.0 00.0 N90 0F
50.? 00.? E; are? 50.0 60.N 05.N 00.0 00.N 006 00.0 006 00.? 9.0. 00.0 00.0. 2..» 9
km... 00... 00.N Emu 00.0 00.0 006 05.0 00.0 00.N 50.N :4 00.N 5.0. 0.6.0 06.0- 00.0 60
no.0 00. _. 00.0 50.N 006 B0... 00.? Sir 00.N 0N.0 60.? 50.N 00.0 9.0- ed 00.0- 00.0 or
0N.0 00. _. 0N.N 00.0 006 {.0 66.0 50.0 N00 00.0 006 00.0 6N.N No.0. 60.N 00.0 00.0 up
06.0 00% 00.0 00.0 006 0N.N 06.0 006 00.0 06.0 50.0 50.0 00.N 60.0- 050 00.0. 5.0 S.
06.0 00. v 05.0 006 00.N 60.N 00.0 ms; 0:” 00.0 05.N 00.N 00.0 0N.0. oﬁw 630. «Na 09
50.? 00.0 0N.0 006 00.0 00.0 00.0 006 N00 00.0 006 F60 66.0 00.0- 00.? 00.0- 556 0
km; 00. F 006 00.0 00.6 00.0 0N.0 006 00.0 00.0 00.0 006 06.0 00.0. N00 00.0- 00. P 0
um; 00... 006 006 006 00.0 006 006 006 6nd 006 006 006 _.0.0. 00.0 00.0 N06 x.
60... 00._. 00.0 006 00.0 0_.._. v0... 06.? 00.N 00.0 06... 00.0 0N.0 00.0 00.0 00.0. 00... 0
N5.0 00.0 00.0 60.0 006 0&6 65.N m E. 06.N {.0 No.0 00.N «0.0 3.0. 00.? 3.0- Fwd m
#0... 00.? 00.N 00.0 00.N 00.0 2N 60.N 9.0 00.9 06.N KN N5.0 00.0- 66.0 00.0- 00.0 6
06... 00.? 5... 0N.F 00.0 \. ﬁr 00.0 50.N had 60;. 00.0 5N 00.0 06.0- 60.N 00.0 {MN m
nmé 00. _. 00.0 006 006 006 006 50.0 50.0 00.0 006 006 60.0 00.0 00.0 00.0 Sum N
00.0 00. _. 00.0 50.? 00.0 00.N 50.0 0N.0 {.0 ~00 00.0 N00 06.0 «0.0- 00.0 00.0 00.5 v
205 CFO—J COO—.1— :"J EMA C”..- CWJ C“...— C":— CM..- CMI- Chl— pMﬂ exp-M pun ”CM 0:3qu0 :3“

 

 

.3331 «59‘ _ um; . 8co0< «cannons... . FEE—£002 005: .2. can...

73

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N04 00. 4 00.6 00.6 00.0 05.0 006 0N.0 {-0 00.0 00.0 00.0- 00.0 00.0 00.0 40.0- 00.0 00
~04 00... 00.0 00.0 06.N 00.0 60.N 00.N 00.0 N0.N 00.0 00.0- 00.0 40.0- 00.0 04.0- N04 60
N04 004 06.0 00.N 054 00.0 00.0 00.N 00.0 00.0 60.0 06.0- 060 00.0- 000 04.0- 004 00
60... 004 «0.0 004 00.N 05.0 006 NN.N 60.0 006 044 04.0- 00.0 00.0- 00.0 06.0- 004 00
N04 00. 4 00.0 50.0 00.0 00.0 06.0 60.0 50.0 00.6 05.0 06.0- 00.0 40.0- 00.0 06.0- 60.0 40
00. 4 00. 4 00.0 00.N 00.0 50.0 00.6 00.0 00.0 40.0 00. 4. 00.0 60.0 00.0- 00.0 00.0- 00.0 00
604 00... 00.0 00.0 N00 KN 0.20 40.0 00.0 004. 00.0 00.0- 004 00.0- 004 00.0. 504 04
504 00... 004 00.0 006 00.0 00.0 00.6 00.6 40.0 00.0 00.0 00.0 00.0. 004 40.0- .3.0 04
N04 004 004 604 00.0 04.0 00.0 44.0 40.0 N0.N 60.0 60.0. 00.0 00.0. 00.0 04.0- 004 5..
064 004 00.0 00.0 00.0 00.0 004 50.N 00.0 06.0 00.0 00.0- 000 00.0- 00.0 40.0- 004 04
N04 004 004 60.0 604. 00.N N04 00.0 006 4.0.0 00.N 06.0- 00.0 66.0- 060 5.0. 00.N 04
604 004 00.0 00.N 00.0 00.0 00.0 064 4.0.0 504 60.0 06.0- 60.0 60.0- 004 40.0- 040 64
n0... 004 00.6 00.6 50.0 00.N 006 04.0 006 06.0 00.0 54.0- 00.0 60.0- 000 00.0- 60.0 04.
004 004 00.0 00.0 60.N N0.N 04.0 04.0 00.0 50.0 60.0 00.0- 004 00.0- 000 60.0- 0.20 04
40.0 004 00.0 04.0 00.N 00.0 00.0 00.0 06.0 004 05.0 00.0- 06.0 00.0- 06.0 00.0- 00.0 4..
064 004 00.0 006 00.0 ..0.N 00.N t4 00.N 00.0 00.6 00.0 mm... 00.0- 04... 00.0- No.0 04.
604 004 50.0 00.0 00.0 60.0 n00 40.0 05.0 00.N 00.0 04.0- 60.0 00.0- 00.0 04.0- 004 0
604 004 05... 04.0 40.0 00.0 40.0 004 504 604 00.N 00.0- 000 00.0. 60.0 06.0- 004 0
B04 004 00.0 00.0 006 00.0 00.0 40.0 00.0 00.0 00.0 40.0- 050 00.0 00.0 06.0- 00.N \-
004 004 00.0 #00 50.0 00.0 04.0 00.0 06.N 604 00.0 06.0- NNN 04.0- 00.0 00.0- 004 0
004 004 40.0 00.0 504 04.0 00.N 00.N 46.0 50.0 00.0 04.0- 000 00.0- 050 04.0- 00.0 0
604 004 66.0 00.0 06.0 00.0 05.0 2.0 00.0 00.0 00.0 60.0- 06.0 46.0- 04.0 00.0- 000 6
N04 004 00.0 46.0 00.6 66.0 00.6 054 00.0 N00 040 66.0- 000 00.0 00.6 04.0- 004 0
004 004 60.N 00.0 60.N 00.0 00.0 004 06.N 00.0 00.0 04.0- 00.0 N60. 60.N 00.0- 004 N
N04 00... 00.0 00.0 00.0 04.0 E20 06.N 00.6 006 00.0 00.0 00.0 04.0- 66.0 00.0- .0.0 4
3°93 C”..— CMJ C”:— CWJ CWJ :"J C”:— CMJ CWJ ”CM «Xv-M R uXCM ”CM “KP—M DESK—Do C3“

 

 

.3503. Ea0< .. £25 - 3co0< E22803... - 8.0.5535. $05... 4.. 20¢...

74

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.0. .00 .00 00.. 00.0 00.0 00.0 0.? 00. 0.? 00. 00
00. 00.0 00. 00.0 00.0 0.0 8.0 ..? 00.0 00? 00.0 00
00.. 00.0 0... 00. 00.0 .00 0.0 0.? 00. 00? .0. 00
00. 0.0 .00 00.0 00.0 00.0 00.0 00? 00.0 00? 0.0 00
.0. 00.. 00.0 00.0 00.0 00.0 00.0 .0? 8.0 8.0 00.0 .0
.0. 00.0 00.0 00. 00.0 00.0 00.0 0.? 00.0 00? 00.0 00
00.. 00.0 .00 00.0 ...0 0.0 00. 0.? 00.0 00? 00.0 0.
00.. 00.0 00.. 00.. 00.0 00.0 00.0 00? 00.0 ..? 0.. 0.
00.. 00.0 00.0 00. 00.. 00.0 00.0 0.? 00.0 00? ... 0.
00.. 00.. 0.0 ...0 0.0 00.0 00.0 00? 00.0 00? 00.. 0.
00.. 0... .00 00.0 00.. 00.0 00.0 00? 00. 00? 00.0 0.
.0. 00.0 00.0 0.. 00.. 00. 0.0 00? 00.0 00.0 00.0 0.
00. 00.0 00.0 00.0 00.0 00.0 00.0 00? 00.0 00? 00.0 0.
00.. 00.0 00. 00.0 00.. 8.. 00.0 0.? 00.0 00? 00.0 0.
00.. 00.0 00.0 .0. .0. 00. 00.0 0.? 00.0 00? 00.0 ..
00.. 00.0 00. 00.. 00.0 00.0 0.0 00? 00.. 00? 00.. 0.
00.0 00.0 .0. 00.0 00.0 .00 0.0 0.? 00.0 00? 0... 0
0.. 00.0 0... 00.0 .00 00. 8.0 00? 00.0 00? 00.0 0
00.. 00.. 00.0 .0. 00.0 00.0 00.. 0.? 00.0 .0? 00.. 0
00.. 00.. 00.0 .0. 00.0 00.0 00.0 00? 00.0 00? 00.0 0
00. 00.0 00.0 00.0 00.0 00.0 00. .0? 00.0 00? 00.0 0
.0. 00.0 .00 0.0 00.0 00.0 00.0 0.? 00.0 00? 0.0 0
.0. 00.0 00.. 00. 00.. 00.. 8.0 ..? 00.0 .0? 00.0 0
00. 00.0 00.0 8.0 8.0 00. 8.0 00? 00.0 0.? 00.. 0
00.. 00.. 00.0 00.0 00.. .00 00.0 .0? 00.0 00? 0.0 .
0.92 0w. a"... 0m. 0m. 5.: pm HM 90.. NH. 25030 :3.

 

 

.3303. Emu< .00-50.... . £cmu< 09.0.... - 83:00.02... 32.... .00 0.0.0...

75

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

464 06.0 00.0 06N N0.N 06.0 00.0 NNN 4~N 04.0 00N 44N N04 00.0- 00N N00- 006 0N
604 00.0 00.0 00.0 404 60.0 0~.0 0~.0 00.0 00.0 00.0 4~.0 00.0 00.0. 04.0 60.0- 600 60
404 ~44 00.4 00.0 00.0 604 00.N 0NN 04.0 60.N 0~N 0~N 00.0 06.0- 60... 00.0- NON 00
60. 4 N00 ~0. 4 60N 00.0 ~0. 4 00.0 4~N 4~N 46.0 00.0 0~.0 00.0 00.0- 000 40.0- ~0.0 mm
004 60.0 N04 00.N 0~4 46.4 00.0 6~N 00.0 40.0 40.0 00.0 00.0 ~0.0- 004 00.0- 0~4 40
04. 4 60.0 00.N 00.0 ~0.0 0~.0 ~0.0 0~.0 0~.0 00.0 ~0.0 No.0 00.0 00.0- 000 0 4 .0- 00. 4 ON
00. 4 00.0 0~.0 00N 00.0 00.N 0~.0 NON ~N. 4 ~0N 0~N 00.0 00.0 00.0 00.0 40.0- 66.0 04
04.0 00.0 044 00.4 0~N ~04 00N 44N 60N ~0.0 00.0 N0.N 60.0 40.0- 00.0 40.0- 00.0 04
60.4 00.N 00.0 00.0 00.4 60.N 0~.0 0N.0 ~44 0~N 00.0 004 00.N 04.0- 0~.0 06.0- 0~.0 ~4
04.4 4~.0 0~.0 06.4 ~04 00.0 604 00.0 00.4 44.0 6~N 00.0 60.0 00.0- 0N.0 40.0- 00.0 04
~0. 4 064 00.0 00.0 00.0 ~00 00.6 00. 4 00.6 00N 00.6 00.0 00.0 00.0 00.0 00.0- 00. 4 04
40. 4 0~.0 06N 00.0 00.0 N0.N 40.0 06N N00 0~.0 66.0 ~0N 60.0 06.0- N0. 4 40.0- 0~. 4 64
044 N04 N00 4~N 464 5.0 00.6 ~0N N04 ~0.0 0N.N 04.0 00.0 40.0- 040 00.0- 00.0 04
604 00. 4 0~.0 00.0 00.0 04.4 00.0 00.0 00.6 00.0 00.6 00.6 04.0 06.0- 60.0 06.0- 000 N4
00.0 004 ~0.0 00N 06.0 00.4 00.0 004 60.4 66.0 ~0N 00.N 06.0 06.0- 0~.0 40.0- 0~4 44
604 004 ~04 40N ~04 00.0 46.0 N0.N 00.N 464 0~.0 40N 6N.0 44.0- 004 00.0- 00N 04
~04 ~44 00.6 60N 00.0 0~4 ~0N 00.0 00.0 00.0 ~0.0 00.0 66.4 40.0- 60.0 06.0- -N 0
06.4 00.4 00.0 ~04 00.0 00.0 004 46.4 ~44 0N.0 00.0 00N 00.0 00.0- N~.0 00.0- 604 0
00. 4 00. 4 4~N 00.0 60.0 06.N 0~.0 ~0. 4 00.0 06N 0~.0 N0.N 66. 4 00.0- 60.0 64.0- 000 ~
00.0 0~.0 0~4 00.N 0~.0 44N 4~N 0~4 00.N 00.6 00.0 00.0 00.0 . ~6.0- 60.0 60.0- 000 0
60.4 4~.0 00.N ~44 00.0 00.0 00N 00.0 00.0 00.0 604 ~N.0 00.0 06.0- N44 00.0- 064 0
064 00.4 60.0 004 06N 00.0 00.0 00.0 ~44 40.0 044 00.4 60.0 N60. 004 00.0- 60N 6
04.4 04.4 «0.0 00.N 064 00.0 00.0 004 00.N 00N N0.N 0~.0 04.0 00.0- 004 ~4.0- 6~4 0
60. 4 0~.0 ~N. 4 06.N ~00 00.N 46.0 00N 00.0 0~.0 04.0 00.N 00.0 ~0.0- 60. 4 00.0- 00. 4 N
N0. 4 00.0 00. 4 46.0 00.6 0~.0 0 4 .0 00. 4 0~. 4 00.N 00.0 06.0 004 00.0- 06.0 ~0.0- ~0.0 4
205 CFO—n— COOI C".— C“.— :”4 End C“:— :"n_ CMJ CMJ CWJ ”CAN- uxcw ”CM“ ax...“ canal—no C3“

 

 

3.3310304: 333.330 - 330< 3x5... - 80.30.32 00...: .04 03.0...

76

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

04.0 00.0 04N 0N.N 04.0 ~04 44N 46.0 00.0 06.N 40N 00.N N00 ~N.0- 0N.0 00.0- 00.0 0N
~0. 4- 06. 4 00.0 00.0 00.0 40.0 06.N 00. 4 00. 4 06.0 00.N N04 6N.0 0N.0- 00.0 ~N.0- 00.0 6N
~0. 4 00.0 0 4N 00.0 ~0N 0 4N ~0.0 00.0 ~0. 4 46.0 N~.0 00.6 00.0 0N.0. 00.0 66.0- 0NN 0N
06. 4 00.0 00. 4 00.N 00.0 0 4N N00 40.0 ~N. 4 ~00 00.N 00.0 66.0 NN.0- N~.0 60.0- 40. 4 NN
00.0- 4~.0 0~.0 6~N 00.N ~0.0 0~. 4 00.0 60.N 00.0 ~N.0 ~0.0 00.N 00.0- 00N 0N.0- N00 4N
0 4.4 00.0 04. 4 00. 4 06. 4 00. 4 00. 4 ~0. 4 NN.N 0N.0 00. 4 ~N. 4 00.0 00.0- 0N.0 0N.0- N04 0N
04.0 00. 4 60.0 0~.0 N0.N 06.N 40N 04.0 00. 4 00.0 00.6 N04 00.0 60.0- 600 06.0- ~6.0 04
00.0- 00.0 4~.0 0N.0 N00 00.N N0.N 00.0 60.0 00.0 0N.0 00.0 00.0 00.0- 60N 66.0- 004 04
60. 4 00.0 40. 4 00. 4 60. 4 NN.N 0N.N 60.N 06.N 60.N 60.0 00.0 00.0 0N.0- 00. 4 04.0- 00N ~4
0 4.0- 000 N00 NN.N N04 00.0 00. 4 00.0 06. 4 N0.N 0~.N NN.N N 4.0 00.0- 0N.0 06.0- 40.0 04
40. 4 60.0 4 4N N00 ~0. 4 00.N 4~.N N0.N NN.N 0N.N 60.0 60.4 60.4 00.0- N0.4 00.0- 00.4 04
N04- 044 EN 0N4 04N 00.N 6~N 44N 00.0 0~.0 60.0 ~0N 04.0 00.0- 00.0 04.0- 0~.0 64
04. 4 EN 46.0 60.N 00.0 00.N 00.0 00. 4 00.0 00.0 00.0 00. 4 06.0 ~6.0- 00.0 N60. 6N4 04
~0. 4 00.0 00.6 N~.0 00.6 04. 4 ~N.0 00.6 N0.0 6~N 00. 4 ~0. 4 00.0 00.0- N4.0 06.0- 044 N4
40. 4 06.N 00.N 46.0 04.0 N04 04N 0~. 4 0 4N ~N.0 00.N 00. 4 00.0 0N.0. N0.0 00.0- 00. 4 44
~04 00.N 604 00.4 N0.N ~N4 04.N 0N.N 06.N 66.0 40.0 00.N 60.0 06.0- 00.0 00.0- 0NN 04
0N. 4- 0~.0 0~. 4 4~N 00.N 00.N 00. 4 0N.N 04. 4 00.N 00.N 46.0 N0.0 00.0- N0. 4 0N.0- NN. 4 0
0N4 0~. 4 N00 ~0.0 00. 4 00.N 04N 06.N 00.N 00.0 0N4 ~0. 4 00.0 4 4.0- 00.0 4N.0- ~6N 0
00.0 00.0 00. 4 N0.N 00. 4 0~.N NN.N 0~.N ~0N 60.N 00.0 00.0 00.0 ~N.0- 60.0 0N.0. N04 ~
~0. 4 00.N 00.N 60.0 00.0 60.N 4~.N 06.N 00. 4 6N4 40N 00.N N00 64.0- 000 N00. 06. 4 0
04.0- 00.0 00. 4 00.N 0N4 00. 4 60.N 00. 4 00. 4 0N.0 00.N N00 00. 4 0N.0- N 4 .0 06.0- 0N.N 0
00.0 00.N 0 4.0 00. 4 06.N 0~.0 00.6 00.0 0~. 4 00.6 ~0.0 00. 4 00.0 00.0- 00.0 00.0 6N.N 6
~0. 4 00.0 N0.N 00.0 N04 00. 4 00. 4 ~ 4.4 06. 4 40.0 00.0 00.0 N44 40.0- N04 40.0- 00. 4 0
N00 0~.0 ~0.0 04.0 00.N 04.4 ~0N 00.0 ~N4 46.0 00.N 00.0 00.0 04.0- N44 .00- 64N N
~0. 4- 00. 4 00.N 00. 4 60.0 06.0 0 4 .0 00. 4 N00 4 4N ~0N 40N 66.0 0N.0- N00 00.0- 40N 4
.2. ..... .0. .0. 0.. .0. .0. .u. .0. .0. ..... .....I 0... 2....

3.331.392. song—.030 .. 330< 09......— - Em.:u:32 00:... .64 «.3...

77

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.. 000 0.0 000 00.0 00.0 .00 00.. 00.0 00.0 00.0 ..? 00.0 00.0- 00.0 00? 00.. 00
.0. .0. 000 00. 00.. 00.0 00.0 0... 00.0 00.. 00.0 .00- 00.0 0.? 0.. 00? 00.0 00
00.. 000 00.0 00.. .00 00.0 00.0 00.. 00.0 00.0 000 00? 00.0 00.0- 00.0 .0? 00.. 00
00.. 00.. 000 00.. 000 00.0 00.0 00.0 000 00.. 00.0 0.? 00.0 00? 00.. 00.0- 00.0 00
00.. 00.. 00. 00.. 00.0 00.0 00.0 00.0 00.0 00.0 000 00? 00.0 00? 00.0 00.0- .00 .0
00.. 00.0 00.0 00.0 00.0 00.. 00.0 00.0 00.. ...0 00.0 00? 00.0 .0? 00.0 00.0- 00.0 00
00.. .0. 000 00.0 00.0 00.0 00.. 00.. 000 0.0 00.0 00? 00.. 0.? 00.0 000- 00.0 0.
00.. 00.0 00. 00.. 00.0 00.0 00.0 00.0 00.0 .0. 0.0 00? 00.0 0.? 0.0 00? 00. 0.
00.. .00 ...0 0... 00.0 00.0 00.0 .0. 00.0 00.0 00.0 .0? 00.0 00.0- 0... 00? 00.. 0.
.0. 00.0 00.0 00. 00.. 0.0 00.. 00.0 00.0 00.0 00.. 0.0- 00.. 00.0- 00.0 0.? 00.. 0.
00.. 00.0 .00 .00 .00 00.0 00.0 .00 00.0 00. 0.0 00? 00.. 00? 00.0 00.0. 00.. 0.
.0. 000 00.0 00.0 00.0 00.0 00.0 .00 00.0 00.. 00.0 0.? 00. .0? 00.. 0.0- 00. 0.
.0. 0.0 00. 00.0 00.. 00.0 00.0 0... 00.0 00.. 00.0 00.0- 00.0 00.? 00.0 00.0- 00.. 0.
00.. 00.0 00.. 0.0 00.0 00.0 00.0 00.0 00.0 00.. 00.0 00.0- 00.0 00.0- 00.0 00? .00 0.
00.. .00 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0.0 00.0 00? 00.0 0.? 00.. 00.0- 00.0 ..
00. 00.0 00.. 0.0 00.0 00.0 00.0 00.0 00.. 00.0 000 0.0- 00.0 0.0- 0... 00.0- 00.. 0.
00.. 00.. 00.0 00.. 00.0 00.0 00. 00.0 00.. 0.. 00.. 00? 00.. 00? 000 00.0- .00 0
00.. .00 00.0 000 .0. 00.0 00.0 00.. 00.0 00.. 000 00? 00.. 00? 00.. 00? 00.. 0
0... 00.. 00.0 00.0 00.0 00.0 00.0 00.. 00.. 00.0 00.0 00.0 00.0 00? 00.. 00? 00.0 0
00.. 000 000 00.0 00.0 00.0 0.0 00. 00.0 00.0 00.0 00? 00.0 00.0- 00.0 00? 00.. 0
0... 00.0 .00 00.. 00.0 00.. .00 00.0 .00 00.0 00.0 00? 00.. 00? 00.0 00? 00.0 0
00.. 0... 00.0 00.. 00.0 00.0 00.. 00.0 00.0 00.0 00.0 00? 00.. 00? 00.0 00.0- 0.0 0
00.. 00.0 00.. 00.. .00 00.. 00.. 00.0 0.0 00.0 00.0 00? 00.. 00? 00.0 .0? 00. 0
00.. 00.. .00 00.0 00.0 00.. 00.0 000 00.0 00.0 00.0 0.? 00.. 00? 00.0 00? .0. 0
00.. 00.0 00.0 00.0 00.0 00.0 00.. 00.0 00.. 0.0 00.0 00.0- 00.0 00? 00.. 00? 00.0 .
0.92 0w. 0w. a”. 0m. 0w. a". 0m. 0m. 5.... n.” WM ”:0. H00. PM. mm. 250.00 :3.

 

 

.3_3310:a0<_= 0632.030 - 330< 30E... - 50.30.35. 00...... .04 030..

78

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

40. 4 00.0 6~N 00.0 00.0 40.0 00.N ~0N 0N4 00.0 60. 4 00.6 66.0 N0.0- .06. 4 40.0- 4N.N 0N
~0. 4 60.N 044 00.N 00.0 60.N 60.N ~04 00.0 00.4 NN.N 6N4 00.4 60.0- 60.N 46.0- 00.0 6N
60. 4 0N4 0N.N ~N.0 ~04 N0.0 ~0N 06.N 00.N 00.N 00.0 00.N 0~.N 0N.0. 60.0 04.0- 00.N 0N
0N. 4 00.0 40.0 ~04 00.0 00.0 04.0 04.0 00.0 N04 ~N.0 4~.N 00.0 00.0- 04.0 00.0- -N NN
604 N00 464 00.0 00.0 00.0 N04 0~4 N04 ~04 0~.N 04.0 00.0 44.0- 0~.N N4.0- 60.N 4N
~0. 4 60.0 60.0 40N 06.0 N00 004 ~N4 00.4 00.4 0N.N ~0.0 0~.0 0N.0- 0~.0 04.0- N~N 0N
04. 4 ~00 ~0.0 00.0 00.N 60.0 0N.0 0~.N 06.4 ~04 ~0.0 04N 60.0 4N.0- 0N.N 64.0- 0 4N 04
~0. 4 00.0 00.0 00.4 NN.N 00.0 06.N ~04 0~4 ~0.0 004 N00 00.0 N0.0- 00.0 40.0- 66.0 04
~0. 4 00 .0 00.0 ~0.0 00.6 00.0 00.0 0N.0 0~.N 40.4 00.0 0~.0 60.0 00.0 00.0 00.0 64.0 ~4
04. 4 4~.0 06.4 0~.0 00.N N00 ~04 004 004 0N.N 0~.N 06.N 00.0 0 4.0- N0.0 06.0- 00.4 04
00. 4 00.4 60.4 00.0 00.4 40.0 N0.N 04.0 ~0N 60.4 0~.N 00.N N0.0 60.0- N04 60.0- ~0N 04
60. 4 00. 4 00.4 0N4 ~0.0 ~0N ~04 00.N 06.N 00.N N04 00.N 06.0 64.0- 664 00.0- N0.0 64
~0. 4 00.N 06.4 6~N 06.4 004 06.N 60.N 00.N 60.N 00.0 40N 04.0 ~0.0- 00.4 00.0- ~06 04
~04 00.0 4~.0 06.N 00.0 04.4 04.0 0~.0 00.6 00.0 00.0 00.6 00.0 06.0- 60.0 ~N.0. 004 N4
604 N04 06.N ~04 00.N 40.0 4~.N 00.0 40N 00.N 44N 604 00.N 40.0- 604 ~0.0- N0.0 44
60. 4 00.4 06.0 0N.0 00.N 00.0 66.0 N~.0 ~0N 0N.N 6N4 N04 N0.N 4 4.0- 004 44.0- ~0.0 04
~0. 4 00. 4 00.6 0N.N 00.0 00. 4 6~.N 0N.N 00.6 N~.0 ~0.0 00.0 66.4 00.0- N00 06.0- 04.0 0
~0. 4 06. 4 004 464 46.4 00.0 004 004 0N4 ~N.0 0~.0 No.0 604 04.0- N0.0 ~0.0- 00.6 0
00. 4 00.0 00.0 ~04 04.0 00.0 00.0 00.0 ~04 04.0 40N 00.0 66.4 ~4.0- 04.4 06.0- 64.N ~
~00 0N. 4 0N. 4 00.0 0~.0 N04 4~.N 00. 4 00.0 00. 4 NN.N 06.0 00.0 NN.0- 00.0 ~6.0- 04N 0
0~.0 60.0 60.0 0~.0 00.0 EN 004 00.N 06.N N00 6~N 00.N 00.0 06.0- N0.0 N0.0- 40N 0
60.4 06.4 00.0 00.0 06.N 40.0 00.0 04.4 00.4 60.0 06.4 04.4 00.0 04.0- 004 60.0- 06N 6
N04 60.0 ~0.0 0~.N ~N.0 00.N 00.0 00.N 0N.N 04. 4 N0.N 60.N 6N.0 00.0- 00. 4 46.0- 00. 4 0
46. 4 N04 N0.N 46.0 66.0 06. 4 ~00 0~.0 00.0 N00 40.0 N0.N 00.0 ~0.0- N~. 4 60.0- N06 N
06. 4 60.0 60.0 ~N.0 00.0 ~0. 4 04.0 06. 4 60.N 00.N 00.0 60.N 00. 4 60.0- 0NN 40.0- 606 4
0.05 CFO—n.- COOVJ CWJ CNJ C”.— C“.— CMI— :"J 2M4 cw.— p—"n— ”CL“ «Kn—M ”a“! «Kn—M sat—no 95¢

 

 

.3503. .0090 _ =02. . 0.0090 09.0... . 50.00000: 000.: .0. 0.000

79

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

86 66. 4 mo. 4 66.4 66.6 606 36 o~6 666 66.4 60.4 466. 86 466. 8.4 066. 66.6 06
~0. 4 666 666 86 60.6 666 66.6 mm...” 36 06.6 8.6 666. 646 666. 606 ~66. 06.4 66
66.4 64.4 66.4 66.4 ~66 646 666 2.6 36 666 666 ~66. 046 646. 66.4 446. 666 mm
~0. 4 666 066 066 646 66.6 666 466 o~. 4 406 066 666- 606 666. 666 ~66- 006 66
66. 4 mm. 4 6m. 4 06. 4 666 666 006 666 66.6 66. 4 66. 4 666 mm. 4 466. o~. 4 646. 666 46
60. 4 606 46. 4 8.4 36 ~66 006 ~06 666 m 46 666 466. 666 ~66. 66.4 666. 06.4 66
mm. 4 .46. 4 R. 4 466 466 666 2.. 4 666 666 40.6 666 666. 606 86- N~. 4 666. 066 64
66.4 066 ~64 046 64.4 066 466 666 666 86 66.4 066. 66.4 86. 66.4 666. $4 64
~0. 4 S. 4 466 NN.N 666 -6 066 666 ~06 666 666 046. 006 066. 66.4 646. 666 5
66. 4 m~6 666 B. 4 .666 006 666 36 36 666 666 ~66. mm. 4 666 on. 4 666. mm. 4 64
~04 666 E6 466 066 606 666 04.6 ~66 066 046 666. 00.4 646. N46 466- 46.4 04
m 4. 4 666 86 2.6 66.6 36 86 666 «66 66. 4 666 666- 666 666. 66. 4 666. 66. 4 64
~0. 4 6m. 4 606 006 646 86 86 666 8. 4 ~66 066 466. 006 666. 666 066. mm. 4 64
66.4 2.6 8.4 60.4 666 666 86 666 406 no.» 666 66.6. 666 646. 666 666. 666 6 4
~04 00.4 ~06 ~06 66.4 046 34 w~6 46.4 666 606 86- 0~4 ~66. 666 666- 3.6 44
~0. 4 06. 4 666 66.4 4 46 466 666 666 60.4 666 66. 4 466. 666 666. 666 666. 64.4 64
~0. 4 66. 4 666 666 666 666 666 mm. 4 466 2.6 86 646. 666 666. 666 666. 066 m
~0. 4 86 606 ~N. 4 2.. 4 ~66 066 mm. 4 2.6 466 666 066. 8. 4 466. 6~. 4 666. 66. 4 6
4m. 4 666 60. 4 066 666 666 066 mm. 4 066 666 86 666. $6 666. 86 666- 066 ~
60.4 m~6 ~06 ~66 86 ~06 66.4 606 ~64 B4 666 066. 06.4 646. 606 666. 66.4 o
06.4 066 666 066 86 666 86 646 466 2.6 66.4 646. 666 666. 66.4 666. 06.4 m
66. 4 9.6 36 6o. 4 466 066 86 666 no. 4 666 on. 4 466. Nm. 4 466. 66. 4 666- 66. 4 6
06. 4 466 no. 4 60. 4 066 66. 4 666 606 66. 4 666 666 666. 64. 4 666. 66. 4 666- R. 4 m
40.4 066 2.6 m~6 86 8.6 066 06.4 3.6 86 «En 446. 606 m 46. 64.4 666- 066 6
66.4 666 64.4 46.4 066 666 2.6 666 606 666 86 466. 66.4 666- 66.4 646- 46.4 4

0.92 a“; 4.“; 4.”.- cwa c"... c".— cma cm; c".— 4>cLM HM 40M “M..."— 4>:Ln_4r 4M...” 0:330 can

$.33. .53 __ 44.25 . 3596 69.5.. . 83:28: 85... .4 4 can»

80

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

60.4 - - ~44 6N.4 46.4 60.0 40.N 06.N NN.N 00.4 04.0 N00 N00- N04 060. 004 N60- 0N4 N>> 0N
40.4 ~0. 4 00.0 N~.0 0N.N 60.N 4~.N 6~.N 0N.N ~N. 4 04. 4 04.N - - 00.0 6N0- 000 0N0- No.4 Nm 6N
04.4 0~.0 04.0 60.0 00.0 0N.0 00.0 00.N 0~.0 6~.N 04.0 ~N.0 - - 60.0 00.0 60.0 00.0- N00 40 0N
~04 - - - - - - 000 60.0 00.N 4~.N 00.0 - - 00.4 400- 00.0 00.0- 004 m6 NN
N04 ~04 N0.N 00.N N0.N 00.N 00.0 N0.N 00.N 0N.0 0~.0 04.N - - 600 00.0 6N.0 NNO- 60.0 40 4N
06. 4 00.0 00.0 00.0 00.0 0N.0 0~. 4 00.0 00.N 600 00.N 46. 4 - - 04.0 0N0- N00 600- ~00 Nm 0N
~0. 4 N0.N 06.4 00. 4 ~0. 4 40. 4 0~.0 00. 4 66.0 00.N 0~. 4 00.4 - - 6N.0 ~00- 00.0 00.0- 40.0 Nw 04
~00 004 40.4 04.N 06.0 00.4 46.0 66.0 00.N 00.0 00.6 00.4 - - 600 06.0- 00.0 440- 00.0 40 04
464 0~.0 00.4 00.N 0N.0 00.N 46.0 0N.N 44.N 04.N 00.4 044 - - 00.0 400- N40 000- 00.0 N0 ~4
60.4 600 06.0 00.N 0~.0 004 00.6 00.0 00.0 4~.N 004 00.0 - - 00.4 40.0. 6N.0 400- 40.4 40 04
60.4 N0.N 60.4 0~.N 0~.0 004 00.0 ~04 000 04.N 00.N N00 - - 60.0 40.0- 0N0 040- N00 Nm 04
0N. 4- ~44 ~N.4 00.N 00.4 ~0.N N0.N 00.6 00.0 00.0 04.0 ~04 - - 60.0 040- N40 ~N.0- 000 N0 64
~04 - - 06.4 00.N 004 00.N 00.0 ~0.N N0.N 0~.N ~0.N 00.4 40.0- 60.0 04.0- 004 40.0- 00.0 N>> 04
604 N04 N00 00.N 00.N 04.4 00.0 04.0 0~.0 ~N.0 00.0 00.N - - N04 000- 000 00.0- ~04 40 N4
604 00.4 0N.N 00.N 0~.0 0N.N 00.N 00.4 06.N 6~.N 0~.0 00.N - - 6N.0 0N0- N00 040. 00.0 40 44
064 - - N00 NN.N 00.0 66.0 4~.N ~44 40.0 06.N N0.N 06.0 ~0.0- N~4 64.0- 00.6 N40. N04 00 04
06.4 - - 00.0 600 4~.N ~00 004 0N.0 00.0 404 N00 60.0 40.0- N~.0 00.0- N0.N 40.0- 00.0 00 0
~04 - - - - - - 40.4 00.0 00.0 00.0 00.0 - - N40 40.0- 000 0N.0. N00 m6 0
00.0 00.4 044 406 60.0 400 00.0 56 ~04 00.N 00.4 44.N - - 00.0 04.0- N~4 000- N60 N0 ~
06. 4- 00. 4 ~04 00. 4 04.N 00.N 00.0 60. 4 06. 4 ~0. 4 00.0 ~00 - - 00.0 00.0- 00. 4 06.0- 000 Nm 0
604 60.0 $6.0 ~04 66.0 40.0 00.0 00.0 00.0 00.N 000 00.0 - - N04 N00- 000 0N0- 6N4 4m 0
40.4 - - 00.4 ~0.N 06.4 00.N 00.N N00 NN.N 44.N 0N.N 044 00.0- 0N0 ~00- N00 000- 00.4 00 6
604 - - 4~.0 00.N 60.4 N~.0 4~.N 00.0 ~00 00.0 00.N N00 660- 004 00.0. 00.0 06.0- 004 N>> 0
00.4 6N4 0~.0 064 00.0 ~N.4 ~00 0N.N 04.N 00.4 04.4 ~0.N - - 66.0 060.. N04 400- 00.4 40 N
60. 4 60.0 400 N00 N~.0 00.0 00.N 00.0 00.6 00.N ~0.N 40.N - - 00.0 00.0 0N.0 06.0- 0N0 Nm 4
205 CPO—n— ...OOFI- CMJ :“n— C”..- :“J CWJ C".— CMJ CM:— C"..- my ex...“ ”CL“ ax...” ”Ch “KC.” Sam—Do Eh. C31

 

 

.353”. 4cou< 030.3230 - Eﬂcusows. was... .04 036...

l

81

Appendix C - Leg Mechanism Case Study - Raw Data

82

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

064 00.04 00.04 04. 44 00.0 00.0 04.NN 00.0- 00.NN 00.~ 00.N 0N
00.0- 0N.N4 00.64 00.N4 00. 4N 0~.0 00.NN 00.0- 00.0N 0~.0 0004 6N
00.0- 00.04 00.04 00.N4 0N.6N 00.~ 0~.NN 000- 06.0N 06.0 000 0N
60.4 00.0 4 0 4. 4N 00.04 00.6N 00.04 00.0N 00.0- 06.NN 000 00.0 NN
0 4. 4- 00.0 0~.04 06.NN 00. 4N 0~.0 00.0N 00.0- 00.NN 0~.0 -0 4N
60.0 00.0 00.04 00.N4 000 04.0 4 00.NN 00.0- 00.0N 0~.0- 0~.~4 0N
N00 000 0~.64 000 00.N 00.0 04.NN 04.0- 0~.NN 00.0 0N.04 04
00. 4- 00.N4 04.04 00.N4 06. 4N 00.0 00.0N 00.0- 00.NN 04.~ 60.0 04
NNO 0N.N4 0N0 00. 44 ON. 44 060 04.0N 00.0- 00.NN 00.6 00.0 ~4
04. 4- 0004 0~.~4 06.64 00.0N 00.0 06.0N 00.N- 00.NN 00.0 0N.0 04
000- 0404 00.NN 0~.N4 06.04 00. 44 00.NN 00.- 0N.0N 000 N60 04
00.0 00.64 00.0 0~.64 00.NN 0~.0 00.NN 00.- 04.NN 0 4.0 0N.0 64
000 00.N4 00.NN 06.44 00.04 0044 00.NN 06.0- 00.NN 000 0N.N 04
~60 00.0 04.04 00.04 00.NN 06.0 0~.0N 00.0- 04.0N 00.0 ~0.04 N4
00. 4- 00.N4 00.04 06.04 00.0 00.0 00.NN 06.- 0~.NN 0~. 4- 06. 44 44
00.0 00.64 00.04 00. 44 0~.6N 00.0 06.0N 04.- 06.0N 00.0 00. 44 04
40.0- 04.64 04.04 00.04 00.6N 00.0 00.0N 0~.0- 0N.NN 00.0 0N.~ 0
040- 00.0 00.0N 000 00.0 00.0 00.NN 0~.0- 00.0N 000 00.0 0
0~.0 00.0 00.04 00.N4 00.0 00.0 00.NN 04.0- 0N.NN 00.0 0N4 ~
0N4 00.6 04.0N 00.04 0~.N4 0~.N4 00.0N 0N.0. 00.NN 000 ~60 0
40.0 00.~ 00.04 00.04 00.04 00.04 04.NN 0~.0- 00.NN 0~.0 N00 0
040- 0~.~ 06.04 00.~4 00.0 4 000 04.0N 04.0- 00.NN 00.6 N00 6
00.0- 00.64 00.04 00.0 0N.0N 06.0 00.NN 06.0- 00.NN 00.0 N00 0
0 4.4 0~.64 0N4 0004 0~.0N 00.04 00.NN 0~.0 00.NN 00.- 0 4.~4 N
0 4. 4- 00.N4 00.64 00.N4 00.NN 00.~ 00.6N 00.0- 0N.0N 00.0 00.0 4
£02 SWI— CWJ CM:— Cwn— C"..— u>CN We.“ ”C.“ 4X6.“ ocan—qno cam—

 

 

.m4_:mem 4:696 30.50“. - 3:094 Eggnog. - 8&5;er 63 64 2%...

83

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

80.0- 08.04 8.44 8.0 08.8. 8.0 8.84 8.04 00.84 8.04 00.8 00.0 00.8 00.0- 00.8 08.0 00.8 8
04.4 8.04 04.8 08.04 08.8 8.04 08.04 00.84 08.04 8.04 8.0 04.84 00.8 00.8 00.8 00.4.. 80.8 8
44.4- 08.84 08.8 04.84 00.8 00.8 00.8 00.0 00.04 8.04 08.04 00.84 8.8 00.4.. 8.8 00.8 8.04 8
40.9 8.8 8... 004.0. 8.04 8.8 04.8 00.04 00.04 00.04 04.8 8.04 00.8 00.0 00.8 8.0. 80.0 8
40.0 00.0 00.04 04.0 8.04 00.04 08.04 00.04 00.84 04.84 04.8 00.04 8.8 08.0. 00.8 8.0 8.04 8
80.4 00.0 8.8 08.04 08.0 00.04 8.8 8.8 8.04 8.8 8.8 8.8 8.8 8.8 8.8 8.0- 00.04 8
40.9 00.8 00.0 8.04 8.04 8.04 00.84 08.04 00.0 04.0 00.0 8.4 8.8 8.0 8.8 00.4 8.04 04
8.4- 00.44 8.44 00.8 8.8 00.04 00.8 8.04 00.04 8.44 8.8 8.0 8.8 8.0- 00.8 8.0 8.0 04
00.4 04.04 00.8 08.04 08.8 00.04 00.8 8.04 04.0 00.4 04.8 00.0 04.8 00.8- 00.8 08.0 8.0 84
80.0. 00.0 8.04 00.8 04.48 00.0 8.84 00.04 00.0 00.04 04.0 8.84 00.8 00.0- 04.8 04.8 84.8 04
40.4- 00.4 00.8 00.8 08.8 00.84 00.8 00.04 00.04 8.8 00.8 08.04 00.8 00.0- 04.8 00.4 8.44 04
00.0. 00.0 00.48 00.8 00.0 8.04 8.8 00.8 8.04 00.04 8.8 08.8 00.8 08.9 00.8 8.0- 000 4.4
8.0.4 8.04 04.8 8.8 00.04 8.04 00.04 00.0 8.04 00.8 00.44 8.04 00.8 8.0 00.8 8.0 00.44 04
00.0 00.0 04.04 00.4.4 00.0 00.04 08.04 00.44 8.0 00.04 08.04 8.84 04.8 8.8- 00.8 8.8 80.4 84
08.9 00.44 08.8 8.04 00.04 08.0 00.04 08.04 08.8 00.04 8.8 00.0 00.8 8.0- 00.8 04.0. 8.8.4 44
80.0. 8.0 00.8 00.8 88 04.8.4 08.04 00.8 8.3 8.48 00.04 00.84 00.8 8.0- 008 04.4- 00.04 04
8.0 08.8 04.0 08.04 8.84 00.84 8.48 00.48 08.04 8.04 00.8 8.44 00.8 00.8 08.8 00.7 00.0 0
8.9 00.8 00.0 00.04 08.44 00.04 08.04 00.04 8.04 04.04 00.04 8.0 04.8 04.8 04.8 8.0- 8.8 0
04.0 8.4 00.48 00.8 00.84 00.84 8.84 00.0 04.8 8.84 8.04 00.44 00.8 00.0. 8.8 00.0 40.0 8
00.4 8.0 8.04 00.8 8.8 08.04 00.04 00.0 8.8 8.8 00.8 00.0 00.8 8.0- 00.8 00.0 00.44 0
00.9 00.8 8.04 00.04 00.04 8.0 00.04 00.04 8.04 00.84 8.04 00.04 04.8 8.0- 08.8 04.0 40.04 0
«0.4- 00.04 00.0 00.8 8.8 00.04 884 08.48 04.04 8.4 8.8 8.04 8.8 00.0 8.8 8.0- 00.04 0
40.0 8.0 8.8 8.8 00.0 08.0 8.84 08.00 08.44 04.04 00.04 8.8 00.8 00.4- 08.8 00.8- 8.8 0
8.0- 00.4. 8.04 00.84 8.8 8.8 00.04 04.4.8 00.0 00.8 00.8 8.0 00.8 8.0. 8.8 00.4 00.8 8
40.9 08.04 8.04 00.84 08.04 00.84 00.04 04.0 8.0 00.04 8.04 8.8 8.8 00.0. 04.8 04.0 8.0 4
0.92 04.44 4.004.. cw... cw. 4.8.... 000.. 4.“; 0w... cm; 0W. 54... R. 40.4.“ «0.4. «MM 250.3 03.

 

 

.3300”. 44.02 _ 000004.030 . 02006 4006000064.. . 400204.00: 60.. .06 0304

84

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.4- 080 00.8 00.04 08.8 8.04 8.04 08.04 00.8 00.8 00.44 00.0 00.8 8.0 00.8 8.8 80.04 8
80.4- 00.8 00.04 00.8 00.8 00.84 00.8 8.8 00.84 04.84 00.0 00.04 08.8 08.0. 8.8 00.0- 8.04 8
80.4 04.8 00.04 00.84 00.84 00.8 00.44 00.8 00.8 8.44 00.8 00.3 00.8 08.0- 00.8 8.8- 008 8
88.0 08.04 00.8 00.04 00.8 8.44 00.04 8.04 8.0 08.0 00.0 00.0 8.8 00.0 00.8 8.0 48.40 8
00.0 00.04 00.8 08.84 04.0 00.04 00.0 00.04 08.0 00.8 00.8 00.8 00.8 04.8- 8.8 04.8- 00.8 48
04.4 00.84 00.0 00.04 04.04 00.48 00.84 00.84 8.04 8.0 04.8 8.4 00.8 00.0- 00.8 8.0- 88.8 8
40.4 8.0 04.04 00.84 00.04 00.04 8.04 00.48 04.84 00.8 00.04 00.0 00.8 00.0. 00.8 04.4. 04.00 04
80.4- 00.8 00.04 04.04 00.8 00.8 00.8 04.8 00.84 00.04 00.8 00.0 00.8 8.0- 00.8 08.0- 8480 04
40.0 00.04 00.0 00.84 00.8 04.04 00.04 00.04 00.04 00.84 08.0 00.0 00.8 00.0 00.8 00.8 80.8 84
8.0 08.04 00.04 8.44 00.04 00.84 04.04 00.04 08.0 00.04 00.04 08.8 00.8 00.0 00.8 00.0 8.8 04
8.4- 08.04 08.0 8.4.4 8.8 00.84 8.04 08.8 04.0 00.8 00.0 00.8 00.8 00.8- 00.8 8.8- 80.84 04
4.0.4- 00.04 8.84 00.8 8.44 00.04 8.04 00.8 00.0 00.84 04.0 08... 8.8 00.0- 00.8- 04.4- 00.00 4.4
00.4 00.8 08.8 04.0 00.0 00.04 00.04 08.04 8.8 00.8 00.4.4 00.0 04.8 8.0 8.8 00.0- 808 04
8.4 00.8 00.04 00.84 00.0 00.0 08.0 8.04 00.44 00.8 00.0 00.0 8.8 00.0- 008 00.0 00.48 84
00.9 00.04 00.8 00.04 08.8 8.0 00.0 04.8 00.84 04.0 00.44 00.8 00.8 08.0. 8.8 8.8- 8.04 44
00.9 8.04 00.0 00.04 00.0 00.0 00.0 00.4.4 00.04 00.8 8.0 04.0 8.8 8.0- 00.8 04.0- 80.04 04
80.9 00.84 04.8 8.8 00.8 8.8 00.8 00.8 04.8 8.0 00.8 08.0 04.8 00.0- 008 00.0. 00.8 0
08.0 00.0 04.8 00.8 04.04 00.04 8.0 04.04 8.04 00.04 00.0 00.4 00.8 8.8- 8.8 00.0- 8.88 0
8.0- 00.04 00.84 00.0 00.4 00.0 04.0 00.04 00.04 00.84 00.8 00.0 04.8 04.? 00.8 04.0- 04.04 8
8.9 00.04 00.04 00.8 04.04 8.04 00.04 8.0 8.04 8.3 04.84 8.0 04.8 00.8 04.8 8.0 8.04 0
004 00.0 00.04 08.04 08.04 04.84 8.8 00.84 00.84 8.04 00.0 00.04 00.8 00.8- 00.8 00.0- 40.8 0
8.4 00.04 00.8 00.04 00.04 00.8 8.48 08.84 00.04 08.04 8.8 00.0 00.8 8.0. 08.8 04.0- 00.8 4.
00.0 00.04 00.04 00.04 00.0 08.04 00.0 8.8 8.0 00.8 00.8 04.44 00.8 00.0- 8.8 8.0- 00.8 0
04.4- 00.04 00.04 00.04 00.04 00.84 00.0 04.04 8.48 00.84 00.04 00.4 00.8 00.0 00.8 8.0- 00.8 8
00.4 8.0 8.04 00.0.4 00.4.4 00.04 00.44 08.04 00.84 04.8 00.8 04.04 00.8 00.0 8.8 8.0 84.8 4
205 04044 0&4; 0”; cw... 0%.. a“; 0W. 0"; cm; 0m; 0.4: pm 40“.". ”:4. W4.- 250400 03.

 

 

.3300”. Eam< = cow-4.2.305 . 8:348 0:34.320... .. 80.22.00: am; .3 03¢...

85

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

40.4- 00.8 00.04 00.04 00.04 00.0 00.0 00.04 00.84 00.04 00.8 00.8- 04.8 04.8 00.8 00.4 00.0 8
00.0 00.0 04.0 8.04 00.0 00.84 00.0 04.04 08.04 00.04 00.8 8.0. 00.8 08.0 00.8 00.4- 04.84 8
04.0- 08.44 08.8 08.04 04.84 00.04 00.84 8.0 00.04 00.0 00.8 08.8 00.8 00.0. 00.8 00.0 00.0 8
00.0 00.0 04.04 00.04 00.8 04.8 00.04 00.0 8.8 8.04 00.8 04.0- 00.8 8.0 00.8 00.0 88.04 8
8.4 00.0 00.0 8.8 00.8 8.04 00.04 00.04 00.48 00.8 00.8 00.8. 00.8 8.8- 00.8 00.0 80.0 48
00.? 00.8 00.04 00.04 00.8 00.8 00.04 00.8 00.8 8.8 00.8 04.8 04.8 00.0. 08.8 8.4- 00.44 8
00.0 00.0 00.0 8.84 00.04 00.04 00.8 00.04 00.8 00.8 00.8 00.0 00.8 00.0. 00.8 00.0 04.0 04
04.0 00.0 00.84 08.84 00.04 00.4.4 8.44 04.04 00.8 00.84 00.8 00.0 04.8 00.0. 00.8 00.0 80.8 04
40.4 00.84 8.8 00.04 00.04 8.8 00.... 8.84 00.84 08.0 00.8 00.0- 00.8 00.8 00.8 00.0 00.0 84
«0.4 00.0 8.04 8.4.4 00.0 00.04 00.04 00.0 8.04 04.0 8.8 8.0- 008 8.9 00.8 04.0 04.0 04
04.4 8.8 00.8 00.44 00.0 8.44 00.04 00.04 00.84 00.84 00.8 00.8 00.8 00.8- 8.8 00.0 00.0 04
00.0 00.0 00.4 8.0 00.84 04.04 8.04 00.04 8.8 08.8 8.8 8.0- 008 00.4.- 008 08.0 8.0 04
00.0 8.84 00.04 04.8 00.8 00.8 00.8 8.8 04.8 00.04 8.8 00.0 8.8 00.0. 00.8 8.0 8.0 04
00.9 8.0 00.84 00.0 00.0 04.8 00.04 04.84 8.8 00.0 8.8 00.8 8.8 00.0- 00.8 00.0 3.0 84
08.0- 8.04 04.8 00.84 8.04 8.84 00.04 00.44 00.4.4 00.0 00.8 00.0. 8.8 8.0- 008 00.8 80.0 44
00.4- 00.0 04.04 8.04 00.8 08.84 8.04 00.84 00.8 00.04 00.8 04.4. 00.8 8.0r 8.8 00.0 80.0 04
00.9 04.04 00.8 8.0 00.8 00.4 8.04 00.44 00.44 00.44 00.8 00.? 00.8 00.0 00.8 04.0 00.04 0
08.0 08.0 00.8 00.04 00.04 00.8 00.8 04.8 00.4 00.0 08.8 8.0r 00.8 8.9 8.8 00.41 8.0 0
8.0 00.44 00.04 04.0 00.0 8.4.4 04.84 00.0 00.8 00.04 00.8 8.4- 00.8 00.8- 00.8 00.8 84.0 8
40.0 00.0 08.0 08.84 00.04 8.04 00.8 00.8 08.8 00.04 00.8 00.0- 08.8 08.8- 00.8 00.0 80.0 0
8.4- 004 8.04 00.8 00.8 8.04 8.04 00.04 00.8 08.84 04.8 00.0 00.8 00.0- 8.8 00.0 00.0 0
80.4 00.04 00.84 00.84 04.8 00.04 00.4.4 804 00.04 00.0 00.8 00.? 04.8 00.0- 00.8 00.0 04.04 4.
00.4 0004 00.4.4 8.0 00.8 00.8 04.04 00.8 08.8 00.44 00.8 04.8 00.8 8.0- 008 04.0 04.0 0
00.0. 00.0 04.4.4 00.04 08.8 08.84 00.8 00.44 04.04 8.0 00.8 8.4- 8.8 00.0- 00.8 00.8 04.0 8
40.0 00.0 00.0 00.04 04.84 00.04 8.0 00.84 00.04 8.44 8.8 04.0 00.8 00.0- 00.8 04.0 04.0 4
I282 0 0 8 0 0 4. 0 8 4 >8 K0 >8 x." 00.4 08.4 258400 03.
=3 00.. =3 00.. 03 00.. 4.0.. 5.. 00.. in 00..! Em Eu 0... 00..

 

 

9.3310592: conga—.35 . 3cou< «cot-82.30:. . 8.022.022 no; .NN 03¢...

86

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

86 no.4 owé nodu 8.3 o 4 .8- nn.n4 o4.m4 no.9 ood nN.N4 no. 44 noon ood- nodw omo nwo nu
00.0 00.0 00.44 08.: 00.08 08.88 04.08 8.0 00.84 00.04 08.04 8.0 00.88 00.0- 00.8 00.8 00.0 8
8.4 00.0 08.4 00.8 00.04 00.84 00.4.4 00.0 00.84 04.0 08.08 00.08 00.88 00.0- 04.88 00.0- 00.00 8
80.0 00.8 00.04 00.08 08.04 8.44 00.84 04.0 04.84 00.44 00.04 00.0 00.88 08.0.. 00.88 8.0 80.0 88
40.4- 00.0 00.0 00.8 04.88 08.04 04.8 00.0 00.8 00.0 00.0 08.0 08.8 00.0. 00.88 00.0 08.0 48
00.9 04.8 8.04 00.04 00.8 08.04 08.48 00.84 8.0 08.0 08.8 00.8 04.8 00.0 08.8 00.9 80.04 08
88.0- 08.4 04.08 00.8 00.4.4 80 00.8 00.04 8.08 00.0 08.88 00.44 8.88 00.8- 00.88 8.0 8.0 04
08.0 00.0 00.0 08.8 08.04 08.8 00.04 08.04 00.0 00.84 00.44 08.04 08.8 00.0 08.8 00.0- 04.04 04
3.6 owd n». 44 oo. 4N n4.o4 oo.n no.4. nvd ov.m4 nvow on.» oo.n4 nwow non oodw no.4? ond \- 4
«so- nv. 4 nwdw nnéu n 4 .o4 o 4 .8 ntww n8..o~ ovdw own 4 nvéw oo.n nudw oo. 4- coda n8..o own 04
36 no.0 o8..o 8.9 8.8 n8..N4 nndm n4..o4 o4.8-4 o8-N4 3.84 9% nn.m~ noo- oN.mw omd 48am n4
vn. 4- nN.m4 oo. 4 no.8 4 non no. 44 no.9 no.3 n8..m4 non 3.84 n 4 .4. no.3 oo.o- oo.~N on. 4- 4~.o #4
86 SN oo.m4 om.» 9.8 8.3 oo.m4 nN.4N oné nv.m4 oo.8.4 nvdw onNN n48- nNNN can undo o4
00.0 04.4 08.04 00.04 00.04 00.44 8.44 00.0 00.48 00.48 00.84 00.8 00.88 8.0 8.88 00.0 8.0 84
4n. 4 owé oo.m4 nvéu nm.v~ noéu om.m~ oo.8- oo.n~ nod on. 4N ooém no.3. n4..o- onNN n8..n- oo. 4m 44
n 4 . 4- own no. 4 nodw nun omN nndw no.o4 o4.8- on.o ov.m4 nv.m4 nn.m~ coo- neon 8.0 86 o4
no.4- nod owe n4.4N n46 oo.o4 oN.o4 no.9 8.8. no.n4 8.4N nmd nndw omn- ooNN n86 8.4.o o
88.0- 00.0 00.0 00.84 08.04 04.04 00.04 00.0 08.8 00.04 00.8 00.04 08.8 8.0- 00.88 00.8 80.0 0
4n. 4 nun ou.~4 n 4.o4 on.NN own no.8... ovom om.n4 oo.4.4 o8..o4 nod noon no.8.- nodw n4.n nod 8-
8o. 4- o 4 .n nod 98.4~ 4 ovd nm.o4 nod nN.n oo.o4 own on.v4 o4..8. ooow no.8.- n8-NN and 4nd o
08.4- 00.0 00.48 00.08 8.04 08.88 00.8 00.84 00.04 08.0 00.48 00.84 00.88 00.8- 00.88 04.0 80.0 0
04.4- 04.04 00.04 8.0 8.08 00.0 00.04 00.48 04.04 00.8 8.04 00.0 08.8 08.? 00.8 00.0 88.8 v
80.4 08.0 00.0 00.84 00.04 00.04 00.48 04.04 08.8 00.44 04.44 00.0 00.8 00.9 00.88 04.0 00.0 0
40.0 00.04 8.84 00.8 00.88 00.0 00.0 8.0 00.04 00.04 00.04 00.0 00.8 08.0- 00.8 00.8 00.0 8
mod. on.m o 4 .v nno nv.o4 90.3.0. nm.o4 oo.N4 nn.o oo. 44 8.04 nod 4 o 4.~N non- nndw oo.o mm. 44 4
0.55 CFO—.4 COO—n. C“.— CMJ :“J C“..— CWJ C".— CMJ CW4 :"J ”c.“ ax...” ”4‘“ exp...“ gnu—DO 63¢
68.3.03. 4cou< _ 442$ - 3:098 4:040:33:— - Sac—0:022 o0.— .nu 038...

87

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

nNo oo.o4 o4.o4 oo.o4 nno no. 44 nx-o 0~.: nv. 44 nNo osdw oo.n- 00.NN no.o ovoN on.n 4nd nN
4nd- ooo no.o4 8. 4w no.o4 on.o4 no.o4 n4. 44 0~.0N ov.o4 no.o-o. oo.o- nnNN oo.n no.oN ooé N4.o VN
vo. 4- own no.o4 oo.o4 ouow o4.n4 oo.N4 no.o 04.4.4 no.3 oodw oo.o- oo.oN no.o oo.oN no.4. on. 44 oN
440.4 oo.o4 oo.n oo.nN ooéw nnéu ovN oo- 4N nu. 4N no.o4 no.oN oo.o- oo.nm nud- nooN on.n- oo- 4 4 mm
ov. 4- no.o own o4.o4 nv.N4 nN.N4 0~.N4 ooé nn.v4 nN-VN ovow oo.o- nwdw nvé oo.oN nnd oo.n 4N
28.0- 86 oN.4 8:04 on. 4N ov.o4 0N.0N no.o nndm nx-NN no.ow on.n- ooNN no.o- on.oN o4.n- on.oo on
oo.o- noé n4.N4 nx-oN no.4N nnN4 00.44 05.n4 00.8-4 on.o4 ovow oo.o- nNoN no.o nNNN ned oo-44 o4
oo.o- oo.N4 oo.n no.o4 nv. 44 04.8.4 and no. 44 nNK4 no.o4 neon oo.n.- nvNN no.o nN.oN 0N.N.- Noé4 o4
cod no. 44 nx-NN n4.o4 ooéw no.4-4 no.o4 no.2 ovém n4.o4 no.oN nné- OYNN oo.o on.ou nod o4. 44 8.4
No.0 0v.» o4.oN oo.o4 noéw nNdN on. 44 n4.o4 on.o4 no. 4N no.N~ ovo- oo.oN on.o ovow no. 4- on.n o4
who nv. 44 no.o4 oo.¢4 nvéw oo. 44 oo.n no.o no. 4N no.oN ovdw n4.®- o4.oN on.n 00.34. no.4 4064 n4
4o.o oo.o4 no.o4 oo.¢4 oo.nN on. 44 an no.o4 nN.\-4 8N4 00.NN 8.7 ooéw own no.oN nvo no.3 v4
No.0 nv.N4 no.3 no.o4 nv.v4 o4.o4 04.4. on. 44 oN.n4 o4.v4 oo.oN no.o ooNN nvo Show no.4. o 4.o4 o4
4Y4 on.n oo.o4 nN.oN no.oN no.o4 on.o4 oméu oo.o4 owow nvNN oo.o- on.oN oo.n on.oN oo.n- no.o N4
mod. on.» no.o4 oo.o4 own no. 4N oo.n no.ow nn.\.4 8N4 no.om owo- nhNN nNo oo.oN no.o n56 44
on.o no.o4 n8.- 44 oo.N4 nv.m4 no.o nsé no.o no.8- ox-n ovdw nvo- nnNN noé 04.NN 0N4 4o.m o4
m4.o- nv. 44 n8: o4.o4 nNoN on.o4 nho on.ow no.o4 no.n4 n4.ow no.7 00.NN nv.» nvNN oo.n ovo o
N~.0 on.44 n8..n4 oo.n4 no.o4 nN.n4 oo.n on.o4 no.oN no.o4 on.ow no.o. on.oN oo.4 oo.oN cod vw. 44 o
no.o on. 44 8.4.4 oo.o4 n4.o4 on.o4 n8..o o4.\-4 oo.o4 no.o4 nnow no.o- on.ow nmé n8..NN oo. 4- vo.o4 -
4v. 4- no.o oo.o4 n4.o4 nde oo.o4 no.o4 no.ow oo.o4 oo.NN noNN no.o- OYNN no.o 00.NN no.4 on.o4 o
04.4 oo.n no.o4 no.N4 nv..8-4 no.o ow.» no.N4 0N.N4 oo.n ovow no.n- nnNN oo.n ooNN omo oo.o4 n
no. 4- no.o o4.o 4 nN.n4 04.N-N o4.o4 coo no.o4 no.o4 on.n-o. oo.oN no.5- OYNN nos n4.oN no.4 no.o e
4o.o- oo.o no.4 8K4 nwdw on.o4 no.ow coo oo. 4N no.4-4 on.oN 040.0 oodw no.o- onNN no. 4- vooo o
on.o oo.n oo.n 4 0N.o4 no. 4N oo.n nvo 04.N4 on.o4 oo.n4 nN.oN ONS- ovNN 05o nN.oN no.o- Nw.o4 N
cod no.o n 4.NN oo.n no.N4 o4.v4 ovo no.ow no.n 4 nn.N4 nndw ood- ooow oo.n nx-Nw no.v on.n 4
O-DE Gun. CNJ C”.— GWJ CWJ :"4 CM:— CMJ :"J «WIM- aXCLM-l ”CM exp—M ”C” by DESK—no C31

 

 

8.3313098 = :35 - £4898 4:02.335 - 83:84.02; o3 .vu 038.4

88

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.0 04.84 00.04 04.04 08.0 00.8 00.88 04.0. 00.88 8.0 84.0 8
4.0.4- 008 04.04 00.04 08.84 08.0 04.88 00.0. 04.8 00.0 00.0 8
00.4 00.0 00.8 00.44 00.04 00.44 00.8 00.0- 08.88 00.0 40.0 8
40.4- 00.04 00.0 00.84 00.04 00.0 00.8 00.8- 00.8 00.0 48.04 88
80.0 8.84 00.44 00.44 08.84 8.8 00.8 04.0- 00.88 04.8 00.0 48
4.00. 00.0 00.04 00.48 08.08 00.0 8.8 00.0- 08.8 08.0- 40.0 08
00.0- 08.04 08.08 08.04 08.44 00.0 08.88 04.8- 00.8 04.0 00.0 04
08.0- 08.04 00.4.4 8.84 00.08 00.0 08.8 00.0- 08.88 00.0 80.0 04
40.4 8.04 08.04 08.0 00.08 8.8 8.8 08.0- 8.8 00.0 08.0 84
00.0- 00... 00.04 00.8 00.48 00.0 08.8 00.0- 04.88 8.0 40.0 04
00.9 00.0 00.84 00.04 04.04 00.8 04.8 8.9 00.88 00.0 00.0 04
80.4- 8.84 8.0 08.04 00.8 00.0 00.8 00.0- 08.88 00.8 00.0 4.4
8.0. 08.44 00.04 08.04 00.4. 08.0 08.8 04.0. 00.88 08.0 00... 04
«0.4 00.4. 00.84 00.88 00.04 08.44 00.88 00.0. 08.8 00.0 00.04 84
00.0 08.84 00.0 00.44 00.44 04.0 00.8 00.0- 08.88 00.4 08.0 44
00.0 04.0 00.84 00.04 00.84 00.04 00.88 08.? 04.8 88 4.0.04 04
00.9 8.04 08.04 8.84 00.08 04.0 08.8 8.0- 00.8 8.0 00.0 0
80.4- 04.8 04.04 8.8 00.88 00.0 08.8 00.0- 00.88 00.0 80.0 0
8.4- 00.04 00.04 08.0 8.0 00.0 00.8 04.8- 00.8 08.0 00.8 8
40.4- 00.0 08.84 08.04 08.04 00.0 08.88 8.0- 08.88 08.8 04.04 0
8.4 8.4.4 08.04 08.04 00.8 08.8 08.8 00.8- 00.88 00.0 00.8 0
80.4- 04.04 08.04 00.84 00.04 04.8 8.88 00.0. 04.88 08.0 80.0 0
04.4 00.8 00.44 00.04 00.04 08.84 04.88 00.0- 00.8 00.0 88.0 0
80.4- 00.84 00.08 00.04 00.84 00.0 00.8 00.8- 00.88 00.0 8.4 8
40.4. 00.04 00.04 04.0 00.04 08.0 00.8 08.? 00.88 00.0 80.04 4
0.05. an... cu. 0m. cm. a". pm... “.8. 4n..." WW4. 25030 03..

 

 

.3303". 4coo< .3950“. - 8:098 .092... - 8.0.4.2005. o3 .nu 038...

89

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4v. 4- cn.c on.o cc.na co.na no.o4 cc.na cc.na cnéa cn.c4 cc.oa no.2 cc.aa cc.o no.oa cc.o- 3. 4o na
aa.c- no.o4 04.4.4 na.o4 no.oa cc.o4 n8..ca cn. 44 nn.o4 n4.a4 no.2 cn.c4 cn.oa cc.o- no.aa cs. 4 oo.o va
n4.-4 on.o na.o nvéa cod ca.o4 nn.o4 no.o4 c4.o4 coda co.o4 no.2 n4.aa cvo- c4.oa cod. 4a.oa oa
Etc cn.o4 no.3 no.o nn.o nn.o no.a4 co.a no.o nn.oa nséa no.o4 caoa no.8.- n 4.aa oo.o oo.o aa
4n.4 on.n n4.c4 na.aa n4.ca oa.2 no.o4 8.4 no. 44 nn.o4 on.o no.o ca.aa nn.o. no.oa on.o 4o.o 4a
oa. 4- 9:. can 4 co.oa no.ca nuo cn.o4 nn.o4 cc.o4 c8..o4 na.ca nha nv.oa c4.o- no.oa co.» no.o ca
o4.c. coé co.o4 co.aa no.3 no.o no.3 ca.c4 no.8. n4. 44 no.2 c4.o no.oa cc.o- ca.oa cc.o NYo o4
cc.o oo.n cc.a4 no.2 co.o4 no.4. c4064 na.o4 na.c4 co.o4 cv.a4 nod c4.aa co.o- no.aa co.a aﬁo o4
and- own oo.o nn. 4a no.aa cho no.2 no.3 on.n4 c4.oa c4.a4 co.a4 ca.oa co.o- no.aa no.8- ao.o4 2
oa.4 co. 44 no.o no.o4 on.o no.c4 nn.o4 ca.o no.o cn.a na.o oo.n co.oa no.o- ccéa co.a on.n4 o4
oo.o- cn.c n52 nn.o non cho cos coo c4.o no.o4 cvca no.o4 co.aa cn.o- cvaa co.n n4.c4 n4
voc- coé co.o4 c4.n4 no.3 coK nv.v4 no.2 cod no.3 no.o4 no.o4 no.oa cn.o- na.aa no.8. 4o.o v4
Voc- co.n4 cn.c4 no.a4 coda no.a4 co.o4 no. 44 nc.a4 no.8. cv.o4 nn.o co.aa n48.- no.aa nod n 4.o o4
ao.o- n4.o co.n4 co.n4 na.o cn.o4 no.aa n4.o4 oo.o n4.o4 coéa co. 44 nvoa no.o- cn.oa oo.n oo.o a4
nv.4 co.n4 nc.c4 no.3 no.a4 no.3 co.o4 coﬁ no.3 cn.o4 n4.oa no.o4 no.oa nv.a- no.aa caK 44.484 44
oo.o nn.o n4.-8.4 no.2 n4.ca nv. 44 no.n4 coé oo.o co.o4 n~.o4 co.4.4 co.aa on.n- nn.oa on.n oo.n c4
o4. 4- no.o4 cv.a4 cc.a4 c4..a4 n4.aa no.o4 c4.a4 no.o no.n4 nn.aa no.o no.oa n¢.a- cn.oa co.4. vn.v4 o
4.8. 4- co.a no.aa no.o4 cc.aa no.a4 no.3 cvéa noN no.o na.oa nvo co.oa co.o- nn.aa no.o on. 44 o
cc.o ca.» no.ca co. 4a co.o4 no.o4 co.o4 n4.o4 nn.a4 cc.o 98.2 no. 44 c4.aa co.o- c¢.oa cvo oo.o4 8.
4n. 4- n4.aa no.a cn. 4a n4.2 n82: co.v4 na.v no.o nv.ca c4.n4 no.a4 c4.aa nn.o no.aa no. 4- na-o o
4n. 4- cc.o4 n4.: no. 4a no.aa n4.a4 cn.c4 c4.o4 co.v4 cn.c4 cn.o4 cc.c4 co.oa on.o. c4.oa non 4o.o n
oo.c- no.4 co.o4 no.4; cc.o 4 c4.a coda no.oa cv.o no.oa n48. c4.aa no.oa cn.v- no.aa no.o on.o w
oo.o- nc.n nn.o4 cc.oa cvéa n~.a4 on.o4 co. 44 nvo c4.n co. 4a 8.2 nn.oa nn.o.- ca.aa on.n nc.c4 o
o4. 4- no. 44 no.4 n4. 4a 912 na.n4 na.n4 no.a4 co.a4 no.o n4.o4 no.8. n4.oa non- nvoa c4.8. 44.c4 a
no.c- no.o coda co.2 co.oa no.2 no.o4 chi nso co.o4 coo no.2 cc.aa cné nn.oa c8..o- vv.o4 4
- . .
a... .4... .0 .0 .0 .0 .0 .0 .0 .0 .0 .0 r... m. 1..- 2......

 

 

.3500: 4coo< _ cow-4.2.8.03 - 3:098 409...... - EmEasoos— oo.— .oa 03.0.4

90

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

No.4 no.N4 co.n no.o4 nn.v4 c5: cn.n cv.n4 cn.c4 cv.44 cn.c4 no.4 no.oN cc.o- nch n4.n- 44km nN
oo.o- oo.4 co.o4 no.3 c4.o nn.o4 no.o4 no.o no.o cod. cn.N4 c4.c4 niwm no.o- cn.oa cod- 4o. 4o «N
oo.4 cc.n chN nn.n nséw no.o cnéw ccéw co.n cn.N nn.n on.o co.oN nos- no.oN c4.o- oodw ow
oc.c cch no.o co.n 4 no.o nn.n no.o cc.n nv.N4 oo.o cc.n cN.N4 co.aa co.n n4.aa nn.n oo.oo NN
4v.c- nc. 4N oo.o nn.o4 n4.: co.n c464 c¢.n4 oo.o cc.v cc.o no. 4 nodw nNo co.oN cud ooNN 4N
Doc nn.4~ cn.N4 cn.c cn.c cc.nN cc.nm cn.c cn.c cc.nN cc.nN cn.c cc.NN no.4 ner 96 3.54 cm
onc nn.o4 n4.oN oo.o co.c4 nvo cv.c4 co.n4 8.: ch. 4N co.» co.o4 no.NN nvé- cc.oN nod- oo.o4 o4
no.c- co.n4 co.n4 cedw n4.o cn. 44 no.o4 co.cN c4.n4 no.o co.o4 nn.o4 cn.oa cv.- co.oN nn.o- oo.oo o4
4v.4 cn.c cn.c co.o4 no.4 co.c4 cn.o4 co.cN co.n4 nc.N4 no.cm co.c4 nn.nu co.o- cc.oN no.o- oc. 4o B4
no.4 nn.o4 no.cN cc.nN cc.nN cn.c cc.nN cc.nN cc.nN cc.nN cc.nN co .o co.oN cc.o chN cc.o o4.oo o4
oo.o 04.NN nN.4 non cwdu coon cod co.n4 nn.c no.n4 na.n4 nn.o4 cc.NN nvo- nvNN no.4... 4o.nN n4
oc.c. cv.c4 no.o co.v4 co.o4 no.cN co.n c5: oo.o no.o4 co.o nv.o4 cn.oa 8.0. nst c¢.c- oo.co v4
04. 4- cn.n 4 cc.nN codw co.v~ nn.o4 no.o4 co.n cv.4 cc.nN cc.nN cc.nN cc.aw cc.o- chw cc.o- oo.oo o4
kc. 4- cc.nN cc.nN cc.nN cc.nN cc.nN cc.nN cc.nN cc.nN cc.nN cc.nN cc.nN cc.v~ cc.o cc.v~ cc.o CNN N 4
oc.c nn.o4 cn.n co.n4 no.3 cc.o4 nn.o4 co. 4N oo.o c4.o nv. 44 n4.o nvNN co.n- nan cn.n- ano 44
ow. 4- co.n coo co.n 4 no.o no.o4 n4.cN co.n4 no.o cn.n oo.o nn.o coon oo.o nsNN on.o coéo c4
oo.o no.o4 nn.n cN.o4 c484 oo.o nv.n4 no.o4 nNéN no.aa co. 44 nvé chN n4.o neon co.c no.oo o
vc. 4 nN. 4N cn.n4 no.n4 co. 44 n4.. 44 nv.N4 no.5 cn.cw cv.o4 nvo nwo cn.oa co.n co.oN no.4. oo.oo o
o4. 4- no.3 cn.n co.n co.o4 cc.o4 co.n4 c4.o4 cn.o4 on.o c4.o4 nc.4 co.oN cN.n co.oN nvc 4n.NN -
nN.c- cn.o4 nvé co.n4 co.n4 cNé4 co.cN no.om cho n4.» no.o no.N cc.oN nn.o nc.NN no.o oo.o o
4v. 4- co.o4 c4. 4 co.n 4 no.o4 co.oN cc. 4N oo.o no.c4 c4.o cc.o nvo nn.ow n 4N- co.NN nv. 4- oo.oN n
oo.o- an c4.o4 0N.NN cn.o4 cvém co.n4 co.o4 oo.o 04.x. cod can nvoN co. 4- coNN n4.n- hvoo 4.
ox-c- co.44 coo nNcN nn.44 nn.o oo.o cn.44 nNoN no.v~ cn.N4 no.o 0~.NN nNK- nNoN c5? nn.4o o
nn.4 cc.nN cn.c n 4.o cn.o4 co.n cn.n 4 n4.o4 nc.c4 cv.c4 on.o no.o co.oN cc.o- ccéw cc.o- oo.oN N
4n. 4 nvé cn.n no.o ccé nn.om cc.nN cn.oa co.ow cn.oa no.vu nn.n cc.vw no.o- no.oN no.o. on. 4N 4
O-DE CPO—h.— CAVOW— C”..— cwn— C”..— C“:- CWJ C"..— CMJ CM.— CWJ ”CM mm ”CL“ MM OCQE—OO 33¢
.333”. .52 __ camcoﬁsm . 852 8.9.5.. . 80.2282 2... .2. «3.84

91

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4n.4- oo.o4 n4.o4 n4.v4 n4.o no.o4 n4.o oo.oo no.ou n4.o ntow o4.~ oo.oo oo.4- oodw oo.4 oo.o nu
oo.o- oo.o o4.o nn.o4 oo.o4 no.o oo.o ooé on.n4 o4.o4 n4.- nvn- nvdm oo.o. nn.ow no.» on.o vm
4o.o nvo nﬁww o4.o4 nvow on. 44 co. 4 n4.w4 no. 44 no. 44 o4.oo on.o. oo.oo oo.o oo.oo oo.o- oo.v4 ow
4n. 4- nvo nod n4.o4 oo.oo no.o4 oo.n4 no.44 oo.o4 oo.o o4.oo nn.n ooéw oo.o- oo.oo oo.n 4n.4. No
5‘4 934 no.o4 3&4 noéw no.o no.om o4.o o4.oo oo.o4 no.aa no.o no.aa on.n- ovdu on.o oo.n 4o
94 no.0 no.o n4.om n~.n4 oo.o4 no.o4 oo.o4 nn.om oo.o oo.oo no. 4- 04.NN on.v- oo.oo no.o n4.o on
9.0 8.04 8.4 00.N4 no.o4 no.o4 8.44 no.o na.n4 oo.n oo.oo nvo- ooéw no.o- oo.oo oo.o oo.o o4
oo.o. oo.o oo.oo nn.n4 nn.o4 no.o4 no.o4 o4.o oo.o ooé n 4.~N on.o. nn.ow ovo- oo.oo oo.4 oo.o o4
34 no.9 no. 44 n4.o4 ov.o4 no.o4 04.44 8.9 no. 4w nvo oo.oo no.4 oo.oo oo.o. cn.oa co.n oo.o 44
oo.o. no.o no. 44 n4. 4N no.om no.o4 oo.4; n4.o ov.o4 no.o4 oo.oo nn.n- no.ow no.o. nn.om no.o on.o4 o4
we. 4- nv.w4 oo.o o 4.o no.n4 on.o oo.n4 oo.o4 oo.nw oo.o4 o4.oo nn. 4 noon oo.n- oo.oo owé no.o n4
5.? on.o no. 4m no.3 oo.o4 8.44 9.3.. noéw cn.44 nod n4.ow no.o coéw no.4- co.o~ non oo.o v4
on.o- no.o4 oo.oo o4.o n4.n4 no. 44 nn.44 2.9 one nn.n no.ou o4.n- no.oN oo.o- coéw oo.o non o4
no.o oo.o n4.o on. 4N no. 4N no. 44 no.9 no.9 on.o4 no.o4 oo.oo on.n o4.oo o 4o- nndo no.o woo N4
o4. 4- oo.o oo.o oo.oo on.o4 8.9 04.04 04. 44 n4..o4 oo.o o4.oo no.o. no.ow non- on.o-m on.o oo.o4 44
8.0- nv. 4 no.o no. 4N no.om oo.o4 na.n4 no.ca nn.ow 04.8- nsdm no. 4 oo.oo n4.o. 04.NN no.o 44o o4
ov. 4- n4. 44 8.0 nn.o4 o4.oo on.n no.o4 no.o4 nv. 4N nn.n4 no.o-m oo.4 no.oN oo.n- oo.oo on.4~ no.o o
no.o oo. 4 no.o oo.oo ooém nv.o4 o4.o4 no. 4w 8.8 nvo nvom n4. 4 04.NN cod. nvNN oo.o no.o o
4n.4 no.o4 no.o no.o oo.o4 o4.o no.o no.n4 no. to 864 ooém on.o nvom oo.o- oo.oo ow.» ovn 4
oo.4 no.n4 no.o4 oo.o4 oﬁvw no.n4 on.o4 oo.n nv.v4 0~.- nudw no.o- nn.ow oo.o. oo.oo oo.n 3.4.4 o
9.0 oo.4 on.4~ o4.oo no.3 n4.o nvd4 oo.o4 8.44 0434 n 4.~N on.o noon no.o. oo.oo oo.n nn.o n
Soo- no.o4 no.8- nn. 44 no.om no.o4 n4.o4 n4.o4 co.nw no.o no.o-u oo.o. nvow nud- non no.o oo.o v
44.4 oo.o oo.o oo.oo nova on.o4 no. 44 ooéw oo.n4 n4.o4 oo.oo no.o- 04.NN oo.o- coow no.o 44o o
ov.4 n4.o on.om no. 44 ov.n4 oo.o4 no.o4 3o co.n oo.n n4.om oo.4- nnNm on.o- no.ow oo.n «No N
3.0 oo.4 oo.4 on.n4 no.o4 oo.oo nn.44 no.o4 no. 4N co.n cn.oa n31 no.oN oo.o- no.aa 04.N oo.o4 4
.35 a”; cw. a? g“; cm; an; em; cm; 54... 9.». 1m”? pm MM RV W“! 83.30 :3.
.333”. an< =_ conconnuuo - 3cmo< 69.5.. - EmenuoE o3 .oN min...

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

nv. 4- c4.o no.ca n 4 cm no.n nN.N4 noéw nc.o4 co.n 4 nvé no.o4 oo.o n».oN c 4 .n- nc.oN no.o 4».o nN
oc.c cn.c n4.o coéw co.oN co.o4 co.c4 n4.. 44 cc. 44 nN.N4 co.o4 c4.o c».oN oo.»- cc.vw co.4. oo.o vN
vn.4 cn.c cc.nN nn.»4 cc.nN oo.o 8.9 no. 44 cn.oN n4.o4 cvdw n».3 nc.NN cn.o- co.aa cc.o oo.o4 oN
oo.c- cn.o no.o4 no.3 nc.o4 co.» c4.o4 co.3 n4.n4 no.»4 c¢.c4 c».o ccéw co.n. no.oN oo.o o4.o NN
o4.4 coo nN.» nn.NN co. 4N no.o4 no.om c4.4N co.c4 co.o4 co.»4 n4.o cc.v~ c».o- co.oN c».4 no.o 4N
oo.o co.a nN.3 c4.o4 noéw cc.o nv.»4 no.o4 n».n4 no.3 co.o4 cn.o no.oN cc.o- nc.oN no.» so cw
oo.c- c4.~ cn.o n4.cN co.o4 co.o4 no.o4 no.o4 nc.»4 co.n4 no.v~ no.» co.oN cc.o. cc.v~ cc.o o4.c4 o4
no.4- no.» n4.o co.44 coNN nc.o no.3 no.o4 no.44 co.o4 co.NN no.» nNNN co.o- no.oN n4.o coo o4
oo. 4- co.4 cn.o noém n».cN co.cN no.cN coé n4.n4 co.o4 c4.o4 cv.» nc.NN co.o- nn.ow co.» oo.o »4
54 co.o4 no.o c4.o4 no.cw c4.o4 no.o4 cN.o nNo cc.o no.» no.n nn.oN not? nodw co.a oo.c4 o4
o4.4 nn.o no.o c4.o no.ow co.n c4.3 cn.44 oo.o co.o4 noéw cc.»4 co.aa oc.c chN no.» on.o n4
oo. 4- co.4 c».n c4.oa no. 4m 8.0m cc.oN no.n4 n».o4 n¢.o4 co.o4 cn.n cn.oa cc.o- co.ow c».4 oo.o 3
o4.c- coé cc.n4 co.o4 ceim n».NN nn.ou nc.o c».o4 c».c4 co.o4 nc.o co.NN n4.n- no.oN no.o o».c4 o4
cc.o cod no.» co.»4 n».cN cvo co.n4 cn.N4 no.o4 no.3 no. 4N no.c4 coNN o4.c- nc.oN oo.o on.o N4
»n.c- no.c n».o nn.»4 c».n4 co.o4 cc.o4 cc. 44 oo.o nn. 4N no.ow co.o4 cv.oN c141 no.oN oo.o on. 44 44
oo.c- c».o no.» co.o4 oo.o no.»4 c4.o4 nc.n no.c4 nN.n4 cc.aw co.N4 n».oN c».N- c4.oa no.o 4o.o c4
o».c co.o4 co.o4 nn.o no.3.- cn.o co.o4 c4.o4 n4.o4 no.o no.o-m cc.44 co.oN no.4- n».oN nn.o oo.» o
o4.c- co.4 no.o4 cc.ow no.n4 co.cN nN.oN no.o nN.c4 no.o4 cn. 4N no.o4 0N.NN c4.»- nn.oN oo.o on o
co.c oo.o cn.o no.n4 co.o4 cc.o nv.o4 co.3 nv.n4 cwo nc.o4 cc.o oo.oo cn.c- c».oN co.» co.c4 »
4o.o nn.o cc.o co.o4 co.o4 nc.o4 cc. 4N no.n4 co.c4 co.n nn.3 co.» nn.ow cc.o- no.oN n4.o 4».o o
oo.c- c».n oo.o n4.NN cc.oN c¢.c4 cc. 44 co.c4 nv.o4 c».c4 cvém nv.» no.oN n».o- nn.om oo.o o4.o n
»c.4- n».N4 nué co.c4 co.3 cn.» no.o4 c¢.c4 nn.3 co.n nv.n4 nvo co.oN no.o- cn.oN nc.» oo.o v
40.4. no.o co.c4 c4.vN nv.44 no.3 c».o4 nn.o nn.o4 n».o4 no.o4 c».n co.oN nn.o- n4.oN co.» cn.o o
o4. 4- n».o nn.o co.cN no.o4 nc.o co.n4 no.o4 no. 44 nc.o c4.oa no.» c».oN co.n- no.oN no.o oc.c4 N
oc.c- no. 44 cn.o no.o4 nn. 4N co.c4 n4.3 co.3 no.o4 nn.n no.3 co.n no.oN co.o- c4.oN cod cn.c4 4
O—OE CwO—u— COOP.— C”.— C”:— C"-— C“.— CWJ C":— e—Ml— CW1— :"A ”CM «Xv—MP ”CM ”CM gal—8 63¢

 

 

.3380. .59.. . :2... . 359.. 3...... . 50.2.2.8... 9.. .3. 2.3

93

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.: 8.2 8.8 8.2 8.8 2.2 8.2 8.2 8.2 8.2 8.8 2.? 8.2 88 8.8 88 8.2 8
88- 2.2 2.2 8.2 2.2 88 8.2 28 8.: 8.2 8.2 2.? 8.8 8.2 8.2 2.2 8.2 8
8.: 8.2 8.2 8.: 8.2 2.2 2.: 8.2 8.2 8.2 8.8 8.2- 8.8 8.2 8.8 8.8 88 8
88 28 8.: 8.2 8.2 8.8 8.2 2.2 8.2 8.2 8.8 88- 8.8 88 8.8 88 88 2
8.: 8.2 8.2 8.2 8.2 2.2 88 8.2 8.2 8.8 8.8 28 8.8 8.2 8.8 8.:- 82 2
8.:- 82 8.8 8.2 2.2 8.2 8.: 8.: 2.2 2.: 2.8 88- 8.2 28 8.2 88 88 8
28 8.: 8.8 2.2 8.8 8.2 8.: 8.2 8.2 8.2 8.8 88- 8.2 28 8.8 88 2.2 2
88 8.2 8.2 8.2 2.8 8.2 8.2 8.2 8.2 8.2 8.8 88- 8.2 88 8.2 8.2- 8.2 2
88 8.2 2.2 8.2 8.2 8.8 88 8.8 2.2 8.2 8.2 88- 8.8 88 8.8 88. 2.2 2
28 8.2 8.2 8.2 8.2 8.2 2.2 8.2 8.8 8.2 8.2 88- 2.2 8.8 8.2 88- 88 2
28 2.2 8.2 88 8.2 2.2 88 8.2 8.2 8.2 8.2 88- 8.8 88 8.8 88- 88 2
88.: 28 88 2.2 88 8.: 2.2 8.2 8.2 8.2 8.8 88- 8.2 88- 8.8 8.? 888 2
5.: 88 8.2 8.2 88 8.2 88 8.: 8.2 8.2 8.2 8.2- 8.2 88 8.2 88- 88 2
88 8.2 8.2 28 8.8 8.2 28 2.2 8.8 8.2 8.2 88- 8.2 28 8.8 88 8.2 2
28 88 8.8 8.2 8.8 8.2 88 8.2 8.2 8.2 2.8 8.? 2.8 88 8.8 88 2.: :
2.:- 82 8.2 2.2 8.8 8.2 88 8.2 2.2 8.2 2.2 88- 8.2 88 8.2 88- 28 2
88 8.: 8.2 8.: 2.2 2.8 8.2 8.2 8.2 8.2 8.2 88- 8.2 88 8.8 88 28 8
8.:- 88 8.2 8.8 88 8.2 8.8 2.2 8.2 88 8.2 88- 8.8 8.2- 8.2 28- 8.8 8
88 88 8.2 8.8 8.2 2.2 8.2 88 8.: 8.2 8.2 88- 2.8 8.2 8.8 2.9 2.2 2
8.: 88 8.2 2.: 2.8 8.8 88 2.2 8.2 8.8 8.2 8.? 2.8 28 2.2 2.: 2.8 8
8.9 2.2 28 8.8 8.2 8.2 8.2 88 88 88 8.8 88. 2.2 28- 8.2 8.:- 82 8
88- 2.2 8.: 8.2 8.. 8.2 2.2 8.2 8.2 2.2 8.2 8.? 8.8 8.? 8.2 88. 8.8 v
28 8.: 8.2 8.2 8.2 8.2 88 8.2 8.2 8.8 8.8 8.9 8.2 8.8 8.8 28 88 8
88.9 88 2.2 8.2 2.2 2.2 8.2 88 8.2 8.2 8.2 8.2- 8.2 88 8.8 88 888 u
8.:- 28 8.2 8.2 8.2 8.2 2.2 8.2 8.2 8.2 8.2 8.? 8.2 28 8.2 28- 8.2 2
292 a”; cw. 8%. cw... cw. a"; cw... cm; a"; H.“ mm. 89.2. In.“ {RF 88.8.... 2.330 :3.

 

 

.2388 .82 __ 82$ . 282 89.5.. - 58.82.82 8.. 88 28»

94

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

888 - - 88.2 2.2 82. 22 8.2 88.2 8882 88.8 88.82 88.2 8.8 88 8.2 88- 8.2 8.2 888 88 8
888- - - - - - - 82.82 2.2 8.2 82.2 88.8 - - 2.8 8.? 8.8 88.8 82.8 8 2.8
2.2 - - 88.8 2.8 2. 8 8.8 88.2 8.2 8.: 8.2 88.2 2.2 888 8.2 88. 8.2 8.2 88 88 8
288 2.2 8... 2.2 8.2 8.8 8.2 8.2 8.2 2.: 88.2 8.8 - - 8.2 88.9 8.2 88 288 22. 2
8.2 - - 8.2 8.2 8.2 88.2 8882 8.8 8.2 88.2 2.2 88.8 2.8- 8.8 8.8 8.8 2.8 288 N>> 8
8. 2- 8.2 8.8 2.2 8.22 82.8 8.2 8.2 8.8 8.2 8.2 2.2 - - 8.8 2.2- 8.8 88.8 28.8 28 8
28.2 2.2 88 8.8 8... 8.8 82.. 8 2.2 8.2 8.2 88.82 2.2 - - 8.2 88 2.8 28 88.2 28 2
288 - - - - - - 88 8.82 88.2 88 8.8 - - 8.8 2.8- 8.2 8.8 8.8 8 2
8. 2- - - - - - - 8882 8.2 8.2 8.8 88 - - 88.2 8.8- 2.2 8.8 28.8 8 22
28.2- - - 8. 22 8.: 8.82 88.8 8.2 8882 88.2 8.82 8.2 2.8 8.2. 88.8 88.2- 8.8 88.8 8.8 88 2
2.2 - - 8.8 8882 88.2 88.8 8.2 2.2 8.2.2 8.8 8.8 88.2 8.2. 8.8 8.8. 8.2 8.2. 88.8 88 2
28. 2- - - - - - - 8.2 8.2 88.2 28 8.8 - - 88.8 8.8. 88.2 88.8 8.2 8 v2
28- 28 8.2 8.2 8.22 8.2 8.2 8.2 8.2 8.2 8.8 8.8 - - 8.8 888- 8.2 8.8 88 :5 2
8.2- - - 8.2 8.2.2 8.2 8.8 8.2 2.2 8.2 88.8 88.8 2.8 28- 8.2 28- 8.2 8.2 22.8 88 2
88- . - 82.8 8.2 2. 22 82.8 88.82 88.8 8.82 82. 8 88 8.2 2.8 8.2 82.8- 8.2 88 88 88 22
28.2- - - - - - - 88.2.2 82. 22 8.2 2. 8 2.8 - - 2.2 8.8. 8.2 888 2.2 8 2
88.8. 8.2 8.8 8.2 8.2 8.2 88.8 8.8 8.2 8.82 2.82 88 - - 8.2 88- 88.2 8.2 2.88 28 8
28.2 - - 8.2 88 88.22 2.2 8.8 88.8 88.2 2.2 88 8.8 88.8 8.8 88.? 88.8 8.8 88 88 8
28. 2- - - - - - - 2.2.2 88. 22 88.8 2.2 8.8 - - 8.8 8.? 2.2 88 88.2 82. 2
28.2. - - - - - - 88 2.82 8.2 8.8 8.8 - - 8.8 88.8- 8.2 2.2 88.8 8 8
28.2 88.2 8.2. 8.2 8.8 8.8 2.8 8.8 8. 22 882 28 2.8 - - 2.2 8.8. 8.2 8.? 28.8 22, 8
28.9 - . 88 8.8 2.2 88.8 82. 22 8.8 8.2 88.2 88.2 2.2 8.2 8.8 2.8. 8.8 8.? 228 N>> 2.
28. 2- - - - - - - 8.8 88.2 8.8 28 8.2 - - 88.2 8.8- 88.2 88.8 88 8 8
28.2- 88 8.22 282 8.2 2.2 8.8 8.2 2.2 8.2 8.2 8.8 - - 2.2 8.2- 8.2 88.8 82.8 28 8
8.2- - - 8.8 8.82 8.8 8.8 8.2. 88.8 88.8 8882 2.2 2.2 2.? 8.2 8.8 88.8 2.2 2.8 88> 2
8.8.2 8282. “.828 8 8.. 8 =3 2 8.. 8 :3 8 :3 2. 8.. 8 8.. 8 8.. 2 :3 pm Um. R “M a”. u.” 88.30 822 2.2.

 

 

.3383". 2cmo< 033-8230 . 82:28:08.2 o3 .48 83.8.8

95

Appendix D - Challenge Mechanism Case Study - Raw Data

96

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

28.2- 2.2 88 8.8 88.8 82.8 88.8- 88.8 8.9 88. 2- 8.: 88
82.9 888 88.8 88.8 88.2. 88.8 8.8. 2.2 888 2.8. 88.22 88
888- 88 88.8 8.8 2.8 88.2. 82.8- 2.9 88- 8.8- .882 88
888. 8.2 88.8 88.8 88.8 8.8 888- 88.2 8.2- 88.8- 8.8 88
82.2 88.2 82.2 8.8 88.8 88.8 8.8 8.8 888. 88.8 28.8 28
2.9 88.8 82.8 8.8 88.8 8.8 8.8- 2.8 8.2- 88.8- 82.8 88
828- 88 88.8 88.8 8.8 8.8 888 88.8 8.2- 88.8- 88.8 82
88.2- 88 8.8 88.2. 82.8 8.2. 2.8- 2.8- 888 2.2- 828 82
288 88.2 8.8 8.2 2.2. 88.8 888 88 88.2 828 88.2 22
2.8.2- 88 8.8 8.8 8.8 88.8 8.? 82.8- 2.2 88.2- 88.8 82
888- 88.2 82-8 888 8.8 82.8 8.8- 88.2 8.2- 88.8- 28.2 82
288- 88.8 2.2 82.8 88.8 8.8 88.8. 8.2 2.2- 28 88.2 2.2
828 8.8 82.8 88.8 8.8 88.8 88.8- 88.8 8.8. 888 88.8 82
8. 2- 888 88 8.8 8.8 8.8 88.8- 888 888 28- 28.2 82
28.2- 8.2 2.2 88.8 88.8 8.8 82.8- 88.? 88 88.2- 82.8 22
828 88.8 28 82.8 88.8 8.8 8.8 88.8 888 88.2 82.8 2
88.9 888 2.8 2.2- 888 8.8 88. 8.8 88.9 8.2- .88 8
28.2- 8.2 2.2 88.2- 88 8.8 88.8- 8.2.. 88 88.2- 82.8 8
888- 88.2 8.8 88.8 888 8.8 8.8- 8.8 8.9 88- 8.8 2
88. 2.2 888 88.8 88.8 8.8 888- 8.2- 8.8. 8.2- 28.82 8
8.2 8.2 88 88.8 2.8 8.8 88.8 88.? 88.8 88- 28.: 8
888 88 88 8.2. 8.8 88.8 88.9 88.2. 88- 88.2- 8.2 v
88.2- 88.8 88.8 2.8 88.8 88.2. 88.2- 8.2.- 88 88.8- 88.8 8
888- 8.2 88.8 2..- 888 88.8 8.2- 88 82.9 8.2- 88.2 8
2.8.2- 8.8 88.2 8.2. 888 88 88.? 8.? 888 88.8- 88.8 2
8.822 cm; and cm; cm; 52: 22.2.” 22...“ RT v.2. 258.80 :3.

 

 

838812598 28.50“. . 8cou< 2:22.380:— . 88.2.2.3: once—.826 .N8 838...

97

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

no.o nn.o 8d on.n 00.N mod oo.o mod 8.». no.8 n _..m 0~.-m 8.2. mwé 88m on.o- 86 mm
nn.o 8.2. océ mvd 02.6 00.N cad nn.o mud no.o mod cn.n mm. 2.- omé 8.8 08.0 mod 8 VN
3.0 0N.N nn.n mad on.o om... owd no.n oﬁm 02-6 00.N mwé no.o- mmd- 8.0 cu. _.- wad mu
nn.o mod oo.n nn.n mum-v oo.n mad 0:8 mud mvd. mod nn.n mud- mod oo.? nn.o so.» NN
3.2 09.0. CY? nn.n ow...“ mod 0~.-w 0N6 nn.n on.n on.n cc.n mmé m _..o- on.o m _..o and 2N
nn.o. 88 8.0 3.0 8N med nn.o on.n nn.n mum-N oo.m oc.c no. 8. cm. 8. ond- oo. 8. mmd ow
oo.o nn.o oo.n oo.n 0N.0 mvé no.n oo.n oo.n 2.4m mud nn.o on.n 08.0- o 2.0 ms; N06 9.
mmd cad 8.0 2.?-m mad oo.o nmé min amt-N mN-m nn.n mum-v 98.0- ms. 7 mvd 9.0- 086? mp
mwd- nn.o on.n omﬁ 8d nn.n oo.o nn.n 0N.0 moé and nn.n 8.8 8.0- mmd- no.o oo.n :
and 8.0 98 98.8. nn.n mmﬁ on.n 8N mud mvd oo.n mud 8N mm... ovd and mac-- or
5.2 owN mmé nn.n 8.0 8.2 on.n 0&6 08.8 no.n no.n nc.o ma... nn.o- and mad on.n 9.
mud. and ovd m .20 no.2 nn.n 09v nn.n ovﬁ mod 96 no.n 8.? ON... 0~.-N oo.n- N93 3
m2. 2. me... owd cod oo.n 3N oo.n no.n nn.n 3.0 mum-v 3.0 CV... and 9.6- ord- vodv 2
mm. 2.- 02.6 nn.n mad no.n mud 9.0 08.0 mad cc.n 02.0 mad 3.0- oo.m- mvd cn.c. and N8
mmd 0~.-F mm.»- ood on.n no.n 01m 8N 00.N 8N. ové mvd 00.8 mm. P- owd and 02.0.. E.
nn.o end 86 oﬁv mad 8N 8.28 no.n 8N mvd 9.6 on.n mvd nn.n 86- 25 K6 or
nn.o mné mad no.n mod mhé 98d nn.n oo.n cn.c cc.n on.n ovd mmé- mmd nn.o- vod m
cc.o on.n 0.2.6 nn.n nn.o and mud 8.8 086 8d 98...” mod mv... no. r. ové on.o- nvd w
nn.o. mm...“ end 9.? no.n omd and mun-2. mad mwd mod cc.o 0N.N. 8.0 8.28- nod on.o -
mm. 8. 8.2 mud no.o on.n mmé on.n oo.n cn.c m ﬁn mvd mud nn.n- And-v- 2.0 on. T end 0
9.6 00.N on.n m ﬁn and oﬁv no.n mod mm... moi muﬁ on.” 8.0- ON.N on.o. 8.2 8.0 n
mud and mmd and mum-m mmd nc.o cn.c no.n mud 0N.0 nn.n 8.2. on... on. F- no.2 86 v
we. 2.- 8.0 00.2 mmé 0N.0 00.N oo.n on.n oc.c 8d ofn no.o oo.n- nn.n- no.o- ood- wad m
96- 2.:-.8 mvé oo.n on.o no... on.n 8N med owé 0N6 omé oo.n- 8N no.o- mr. F- vm. 3 IMI:
mud 0N.N mod mwé end 91.... 09v oo.n on.n mvd on.n mvé mm. _.- mad nn.o on.o. 52.0 2.
0.93 22.”; can—n. 22“.. 2.”... :2...— cwq 2."... cr cw.— cm.. 22".. We“ ”mm. may ”2...." 0:32.30 :3.

 

 

2.3881252. 58:82.33 - 2:892. 2:82.822:— . 82:82.08: 89.2—82.0 .88 838.»

98

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

nn.o. mwé 8.6 06.6 nn.n on.n on.n mum-6 mN6 on.n 0N6 3.0 and no.8 nn.o- on.n 56 mm
5.0 m6.N on.n on.n oﬁm no.n Q66 8N no.2. 06.0 on.n oo.o mﬁw 06.0 06d 0~.-o No.2- 6N
5.8 mﬁm on.n 8.6 066 oﬁm oo.o oﬁo 06d 8.0 066 on.n and mad- 0N.0 00.0 and mm
mm. _.- CNN DNA» 9.6 m66 ow6 no.6 on.n 06.2.” 86 nn.n mad mo.m- no.o- m6.o- mmN- wad Nu
mud 8.6 mmé 36 mod mud 2.6 mﬁm 3.6 86 mﬁN 06d and- 0&8 oﬁ T 00.N mm. 5. 2N
and- oﬁm nn.n nn.n oo.o 0N6 8.6 m ﬁm no.8 on.o 06.m 00.0 8.0 06.0 one no. T «6.2- on
20... on.n mod om.m m6.m om6 mud nn.n 06d mmé om6 on.n and 0~.-o m6.o no.n mmﬁ m2
and mad mod 3.6 8.66 mus-.8 8.2 oo6 0N6 mﬁ6 no.2 mad 08.8 on.o mud oo. w- 5.» 8
6a.? oo.n nn.n on.o no.2 no.o mod oo.n mﬁo nn.n on.o mum-N on.n 0~.-o- omd 8.0 26.8 t
N~.0 oﬁr nn.o no.6 mud oo.n nn.o mwd 0~.-m on.n on.o 0N6 omé nn.o- oﬁo 0~.N NB. 3 or
Oﬁ F- r66 06d msé on.n 26 on.o no.6 no.o no.6 mad 9.6 00.0 nn.o no.2 0~.N «m2. m _.
mod on.n no.2 oo.o on.n 06.6 86 no.6 mud oﬁm oﬁm 0N.0 omé and 86 on.o. 6M2- 6..
26.8 and on.n omé mod 86 om6 on.n nn.n nn.n mud oo.—- on.n- 8.6 06.6- 086 on.o m.-
mmd- 8.8 00.6 mad no.6 omd 8.9 06.6 on.n on.n oﬁm mud on.n- oﬁo owd- mNN- 00.0.. N.-
mo.o- no.n 8.6 oo.n 00.6 mud.- mw6 omé 2.2. 02-6 3.6 on.o nn.o oﬁo on.o no.8 on.o :
2.0- mwd no.6 mud oﬁN m6.m 8.6 mN6 mwd 8.2 066 and end 8.2m- mm.m- no.n- 8N6? or
6m.o oo.n on.n om6 on... 86 8.6 mm6 8.6 mﬁo mad and 06.0 mum-o nn.o mud. oo.ow m
and on.n mad nn.n 3.0 00.6 8.0 omé mﬁ6 0N.0 nn.n O66 00.N oﬁo- 0N.N 06.8 8d 9
~N.0 m6.m 36 m ﬁn 0N6 nn.n on.n on.n om6 86 06.6 m6.m omé 0N.N nn.n 09m 86 -
wwd no.2 mad o ﬁ6 0N6 owd m66 on.n mum-m nn.n 8.8.8.8 on.n 0N.0- omd nc.o- mmd- @062. 0
0N. T 066 mmé m ﬁm 06.8 mﬁm on.o on.n no.n 8.6.2 00.6 no.o mmé omd- omd 02-.— 86 m
mmd mm6 8.8 on.n oo.n nn.o on.o no.8 m6é nn.o no.n mod m6.o 06.0 nn.o 8.0 mod 6
2b... on6 cod nn.n 2.6 nn.n on.n O66 omd 8N om6 2.6 no.o no.o omé oo. F- 8. I. m
oﬁp 8N on.n omd on.n on.n om6 mﬁm oﬁN m6.m oo.n mm6 and. mud nn.o. oﬁN 06.2- N
60. F- 2.6 $66 06.8». mmN on.n 2.2 mw6 8.8 no.6 no.n mum-6 oﬁm- 8.0- 86.- mu. 8. 60.6.. _.
205 EFG—.4 600?:— e—“n— C“..— CWJ C“.— CWJ :"n— CWJ CW1— CNJ m“! ”a.“ ”CM “M ocak—DO 63¢

 

 

.8388”. 2:898 = 2208:2385 - 3:898 22.22.288.822. - 88.22.82.085. 89.2.9.0 .68 28.8...

99

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.0 00.0 00.0 00.0 00.0 00.0 0.0 00.0 00.0 0.... 00.0 0~.. 00... 00.0- 0~.? 00.0 0.0 00
00.0 00.0 00.0 0.0 00.N 00.. 00... 00.0 00.0 00.0 0...- 0~.0 00.0 00..- 00.0 00.N 00.0 «N
«0.0 00.0 00.0 00.0 00.0 00... 00.0 00.0 00.0 0.0 0~.0. 0: 00.? 00..: 00.? 00.. ~0.N. mm
00.0 00.0 0~.0 00.0 0.0 0~.N 00.0 00.N 0~.0 00.0 00.0 0... 00..- 00.0. 00.0- 0~.N «.0. «N
0...- 00.. 0~.N 0~.0 0~.0 0.3 00.0 00.0 00... 00... 00.0- 0... 00.0. 0...: 00... 00.0- 00.0 .N
00.0 00.0 00.0 00.0 00... 00.0 00... 0.0 00.0 0.0 00.. 00.0 00.0- 00.0 00..- 00.0 .00 00
0.0. 0~.N 00.0 00.0 00.0 03. 0~.0 00.0 00... 00.0 00..- 0.0 00.0. 00.0- 00.0. 00.0- 00.0. 0.
«0.0 00.0 00.0 0.0 0~.0 00.0 00.0 00.0 0.0 0.0.... 00.0 0.0 0... 0~.0 00.0 0~.0 0: 0.
00.0 00.. 00.0 0.6. 00.0 0.0 00.. 0~.... 0.0 0: 00.0 00.0 00.0 0.0 00.0 00.9 0~.0 t
00.0. 0~.. 00.. 00... 00.0 03. 0~.0 0~.. 00.0 00.0 00..- 00.. 00... 0.0 00.0 00.0- 00.0 0.
0~. .- 0.0 00.0 00.0 00... 00... 00.0 00.0 0~.0 00.0 0...: 0~.0- 00.N 00.0 00.9 00... .0. 0.
2.9 00.0 03 00.0 00.0 0... 0.20 0.... 00.0 00... 8.0- 00.0 0~.0 00. .- 00.0 00..- 00.0. v.
0.0. 00.0 00.0 00.0 00.0 00.0 00... 0... 00.0 00.0 00.0- 03- 0~.0 0.0- 00.0 00.0- 00.0 0.
00.. 00.0 00.0 0.0 00.0 00.0 8.0 00.0 00.0 0.0 0v... 0~.. 00.0 00.... 0~.N 00.0 00.0 N.
0...- 00.. 00.N 00.0 00.0 00.0 0~.0 00.0 00.0 00.0 00.... 00.0 0.0. 0~.0- 0~.0 00.0- 0.0.0. ..
00.0 00.0 00.0 00.0 00.0 00... 00.. 00.0 00... 00... 00.. 00.0 0...- 03. 00.0 00.N 00.N. 0.
00.0. 00.0 00.0 00.0 00.0 00.0 0.... 0: 0.... 00.0 059 8.0 00..- 0...- 00.0 00..- ~0.N. 0
00.0 0.... 00.0 00.0 0.0 00.0 00.. 0.0 0~.0 00.0 0.0. 0... 00.0. 00.0. 00.0- 0.0 .00 0
00.0- 00.. 00... 00.0 00.0 00... 00.0 00.0 00.0 0.0 00.0- 0.0 0... 00.9 00.0. 00.. ~60 .
0... 00.0 00.. 0~.0 00... 00.0 00.0 00.. 00.0 0~.N 0N.N 8.0- 00.0- 0.6 00.0 0.6. 0.0 0
0...- 0~.N 0.... 0.0 00.0 00.. 00.N 00.. 00.0 0.0 0~.0- 02 00.0. 0~.N- 00..- 00.0. 00.0. 0
00.9 00.0 00.0 00... 8.0 00+ 00... 03. 00... 00.0 00.N- 00.0 0.0- 00.0 00.0- 0.0- 00.0. v
00.0- 00.0 00.. 00.0 0.0 00.N 0....” 00.0 00.0 00.0 0.....- 00.0 00.0. 00.0 0.6.- 00.0. 00.0. 0
.0..- 00.N 00.. 00.0 00.0 00.0 0~.0 00.. 00.0 00.0 00.? 0~.... 00.0 00.9 00.0- 0.0 00.0 N
3.0 0... 00.0 00.0 00.0 8.0. 00.0 00.0 00.0 00... 00.N. 00.N 0~.0 00.? 00.0- 00.0. 3.0. .
«.05 an; 0“; 0%.. cw; 0“; 0...... cm; cm; 5.... 9M .0...“ Wm! WM .0...“ w... 250.3 00¢

 

 

9.3310592: comcasaoam . Ecau< 0:00:33... . 8.0.5502 09.2.25 .3 03¢...

100

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

60.0 00.0 00.0 00.0 00.N 00.6 00.0 06. e 00.N 00.6 00.0 0N.0 09.0 00.0 00.0 00.N 0N.0 0w
..0.0 00.0 00.0 00.0 0N6 00.0 00.6 006 00.0 00.0 00.0 09.0 00.0. 0~.0 00.N 05.0 00.0 6N
00 .0- 0N.N 05,—. 056 00.0 00.N 00.N 00.0 06.0 00.N 00.0 0N.0 00.6. 00.6 0 0.. 00.N 00. E. 0N
00.0. 00.0 00. r 00.0 00.6 06.0 00.0 06.0 0N6 00.N 00.N 06.6 0w6. 066 00.0 0rd 0~.N? mm
00.0. 00.0 00.0 00.6 00.N 05.0 00.0 00.N 00.0 00.0 0N6 00.6 00.N- 00.0 0N6. 006. as. Z. 2..
56.0 00.0 00.N 0...6 00 .0 06.6 06.0 00.N 00.0 00.0 00.6 00.0 06... 00.0- 00.0 00.N 56.0 on
0N.0 00. r 00.0 00.0 06.0 0N.N 006 00.0 00.6 06.N 0.20 00.N 00. F- 00.6 00.0- 00.0 00.x... 2
00.0- 0N... 00.0 0N.0 00.6 00.N 00.0 00.0 05.0 00.N 00.N 00.0 00.0. 05.0. 00.6. 00; 3.00 00
66.0 00.0 0N.0 00.6 00.0 00.0 00.0 06.0 00.6 00... 00.0 0...6 00.N 00.0 0N.0. 00.N 006 t.
00.0. 00.0 00.0 0N6 00.0 0N.0 00.0 0~.N 00.0 0~.0 00.0 00.0 006. 0~.... 00.0- 00. T 0N.N.. 0..
0 r. .3 00.0 00.0 00.0 00.0 00.0 00.0 00.6 0N.0 00.0 056 0N.0 00.0- 00; 0N.N- 00.N- 00.0w 0?
00.0- 00.6 0~.0 00.0 006 06.6 00.0 05.? 0N.N 00.0 0~.0 0~.N 00.0 00.0- 00.0. 0N6 00.00 60
00.0 0N.0 00.0 00.N 00.6 00.0 0...0 00.6 0~.0 00.N 00.0 00.0 00.0 00.0. 0~.0. 06.0. 650.. 0..
00.0 00.0 00.0 00.6 00.6 0N.0 00.6 06. r 00. r 06.0 00.N 09.0 00.N- 00.N 00.N 00.0 00.0 up
60.0. 00.N 06.N 0N.0 00.0 00.0 00.0 00.N 00.0 06.0 06.0 0N.N 00.0. 006 0N6. 9N 60.00 5.
~00. 00.N 00.0 00.0 00.0 06.0 0~.N 0N.0 00.0 0N.N 00.0 00.6 00.6. 00.6. 0N. ..- 0N.N- 60.0.. 0..
_.6.0. 00.N 00.0 00.0 06.0 00.6 00.0 0N.N 00.0 on... 00.0 06.N 00.0- 00.N 00.N- 0N. P- 09$ 0
60. ..- 0~.0 00.N 00.N 00.0 00.N 05.6 0N6 06.6 00.N 06.0 00.0 00. T 00.0. 00.0 06.N- 0N.0w 0
60... 00.0 00.0 00.0 06.N 00.0 00.0 00.N 00.N 0 0.0 00.6 00.N 00.0. 06.6 00.N- 00.0- N60 n
00.0. 00.0 00.0 00.0 00.0 0.20 00.0 00.N 05.0 2.0 00.0 06.6 00.N- 0N.0. 00.0 00.0- 60.N. 0
00.0- 00.0 00.0 00.0 0N.0 00.? 00.0 00.0 00.6 06. —. 0P0 03.6 0~.N 00.0 06.6 0 0.0. 50.3 0
60. F- 00.0 00.0 00.0 00.0 00.0 00.6 00.6 00.0 00.N 006 06.0 00.6 00.6- 0N6 00.0 50.09 6
0w... 9.0 00,—. 0~.0 06.0 00.0 0...0 00.N 0~.N 00.0 00.0 0N6 00...- 00.0 0N. T 0N.0- 00.6 0
00.0- 00.0 06.0 00.0 00.6 00.0 06.0 0~.0 06.0 00.0 00.0 00.0 006. 00.0 06. F. 00.6. 60.0? N
00.0. 00.. 00.6 00.6 00.N 00.0 00.6 06.N 00. _. 00.6 0.20 00.6 00.0- 0N6. 0N... 06.0. 60.0 F
265 CFO—n.— COOFJ EMI— CWJ Cﬂn— C“:- CWJ :"J C”.— CWI— Cr ”C” axe.“ ”CM ““1 Ocalg 63¢

 

 

.3500”. E00< . £02. . £c00< 0~.0—0:30.00. . 80.205002 00:0..030

.00 0.00h

101

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N~.0 00.0 0F.N 00.0 00.6 0~.0 0N.0 00.0 0N.0 00.0 0F.F 0 F0 00.0 00.0 00.0- 00.0- 60.2 00
F00. 00. F 00.0 00.0 9.0 0 F6 00.0 00.0 05.6 00.0 00.N- 06.N 00. F- 00. F- 00.N- 00. F- ~N.0F 60
00.0- 00.0 00.0 05.0 06. F 00.6 00.0 00.N 00.N 00.0 00.0 00.N 00.6 0~.N 00.0 00. F 00.NF 0m
00. F- 0~.N 00.N 0F.0 00.0 06.N 00.6 00.0 00.6 006 006- 00.0. 00.0- 00.0- 00.F 00.0. 00.0F NN
F0. F- 0N.F 00.N 00.6 00.0 00.0 05.0 00.N 00.6 00.0 00.6- 050. 00.0 00. F. 0...0 00.6- 00.0.. FN
00.0- 00. F 00. F 06.0 06.6 00.N 00.0 00.0 06.N 0N.0 06.6. 06. F 00.0 00.0. 00.0 0~.0- 00.0 cm
00.0- 00.0 0F.~ 0F.0 00.0 00.6 2.0 0N.0 00.0 00.6 0~..F 0N.0 00.0 06.F 0~. F- 00.0 0N.0 0F
60.0 00.0 00.0 00.0 0N.0 0 F.0 06.0 006 26 00.0 9.0. 00.N 00.6 00.0- 00.N 06.0 00K 0F
50. F- 06.F 00.N 00.0 00.0 0 F0 00.0 06.F 00.0 00.0 00.0- 00.0- 00.0 00. F- 0~.0 00. F- N00 NF
06. F- 0N. F 0F.N 05.6 00.0 00.6 00.6 0...0 00.N 00.0 06.6- 00.N- 06.0 0F. F- 00.0 00.N- 00.0 0F
F0.F- 00.6 06.N 0~.0 00.0 00.0 00.0 00.0 06.0 00.0 00.6- 00.0- 0F.F- 00.6- 00.0- 00.6. 60...F 0..
F0. F 0~.0 00.0 006 00.N 00.N 00.N 00.N 06.0 06.N 06.N 05.6 00.N- 00.0 00.N- 0~.0 00.0 6F
05.0- 00.0 9.0 06.6 0~.0 006 00.0 00.6 00.6 00.0 00. F- 0F.N 0... F- 00.N- 00.N- 00.N- N00 0F
00.0- 000 00.F 006 00.0 006 00.N 36 00.6 00.6 0N.0 0N. F- 0N.F 00.0- 0F.F 00.0- 00.0F N F
06. F- 00.N 00.N 0N.0 0N.0 00.6 00.6 00.6 0N6 05.0 00.0- 00.0- 00.0- 00.6- 0F.0 00.N- 06.0F FF
N~.0- 00. F 0 F.N 00.N 00.0 00.0 06.0 0». F 0~.F 00.0 00. F- 00.0 00.0- 0N.0 06.N- 0...0 006F 0F
00.0- 00. F 0.0.6 006 00.0 0N6 00.6 00.6 0N.0 0N6 00. F- 00.F 00.0- 0F6- 06.6- 00.N- 0F. FF 0
00.0- 00.N 00.N 06.0 00.6 00.6 06.N 06.0 0 F6 05.0 00.0- 0~..F 00.0. 00.0 00. F- 00. F- «0.0 0
00.0 00. F 00.6 00.0 06.F 0 F0 0N.N 00.0 0N.0 00.0 00.N- 00.0 00.0- 00. F- 00.N- 06.0- 60. FF \-
000- 0N.0 00.0 06.6 00.N 9.0 05.6 00.0 00.0 00.0 00.0- 00.0- 06.N 06.F 00.6 00. F- 00.0F 0
50.0 0~.0 0 F.F 056 056 06.0 00.0 06.6 00.0 06.0 00.0 00.N 06.0. 0N.F 00.N- 00.6 F00 0
00.0 00.N 0N.F 0F6 00.0 2.0 00.0 00. F 2.0 00.0 0~.N 00.6 00. F- 00.N 0 FN- 00.N- ~0.NF 6
0N. F- 00.0 00.N 00.0 00.0 0N6 0~.0 05.0 9.0 0N.0 00.N. 06.N 0F.0 00.N- 00.N 00.N- N6. FF 0
60. F- 00.N 00.F 00.0 00.0 00.0 00.0 00.F 0F6 00.6 0~.6- 00.0- 9.0. 00.N- 00. F- 05.6- 0N.0F N
50. F- 00. F 0N.0 00.6 00.0 05.0 00.6 00. F 0~.0 00.6 00.0- 00.6- 00. F 00. F- 00.0- 00.0- 65.0 F
205 EMI— CMJ C”.— C“.— CW4 CHI— CMJ CM.— CHI— |~>¢M WM bur «Xv—M ml «ﬂ OCQEEO pun—K

 

 

.00.—.00”. E00< ._ .002. - 00:00< 0000:0000... . 80.003005. 00:0..050 K0 030k

102

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.0 00.0 00.0 0.... 00.0 00.0 0..... 00.0 0.0. 00.0 00.0 00
.00 00.0 00... 00.0 00.. 0...... 00.. 0.0 00.0 0... 00.0 0~.
3.0 00.0 00.. 0.0 00... 00.0 00.. 0.0 00.0 00.0 t... 00
3.9 00.. 0.... 00... 00.0 00.0 00.0. 00.. 00.0 00.0. 00.0 00
.00 00.0 00.0 0.0 0.00 0~.0 00... 00.0 00.0- 00.0 00.0 .0
0~.? 00.0 00.0 00.0 00... 0.0 00.0 00.0 00.0- 00.? 0..... 00.
0.0 00.0 00.0 00.0 0.0 00.0 00... 00.... 00.0 00.0 00.0. 0.
3.0. 00.0 00... 00.0 00.0 0.... 00.0- 00.0. 00.0- 00.0. 00.0. 0.
0....- 00.0 00.. 0.0 00.0 00.0 00.0- 00.0 00.0 00..- 00.0. t
3.0- 00.0 0.0 00... 8.0 00... 00.... 00.0. 0... 00.0- .0... 0.
00.0. 00.. 00.0 00.0 00.0 00.0 00..- 00.0 0....- 00.0. 00.0. 0.
.0.. 0... 00.0 00... 00.0 00.0 00.0 00... 0.0 00.. 00.0 v.
00.0. 00.. 00.0 00.. 00.0 00.0 0.0. 0.0 0...- 00.0 00... 0.
00.0. 00.0 00... 00.0 00.. 00.0 00.0. 00.0. 0... 00.... .0... 0.
00.0. 00.. 0.0 0.... 00.0 00.0 00.0.. 0... 00.0. 000- 00... ..
00.0. 00.0 00.0 00... 0~.... 00.0 0.0- 0.0 00.0. 00.0 3.0 0.
00.. 00.. 0.0 00... 00.0 00.0 00... 00.. 00. .- 00.0 .00 0
0.0. 00.0 00.0 00... 00... 00.0 00.0- 0.0 00.0 00.? 0.0. 0
.v..- 0... 00.0 00.0 00.0 00.0 00.... 00.... 00.0 00.0. 0.00 .
2.9 00.0 00.0 00... 0.... 00.0 00.? 0... 00..- 0...- 0..... 0
00.0 00.. 00.0 0.0 0:. 00... 00.0 0.0 00.0. 0.0 :00 0
00.0. 8.0 0.0 00... 00.0 0.... 0.0- 00.0 0..... 00..- 0.... ..
00.0. 00.0 0.... 00... 00... 00.0 00.0. 0...... 00.0 00.0- 00.0 0
00.0 00.. 00.0 00.0 0.0 00.0 0.00 0.0 00.0 00.. t0 0
.00. 00.0 00.0 00.0 00.0 00.0 00.0- 00.0. 0...- 00.. 00.0. .
0.0.2 cm. a". 0m. cm. a". 9.0.. 00...”... 9%... a...” 250.00 .50

 

 

.00—:03. 0:00< .00-So“. - 00:00< 000...... - 50.000005. 00:0..0zo .00. 0.00...

103

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00. F- 00.0 00.F 00.N 00.0 00.F 00.N 0~.N 00.6 00.N 00.6 0F.0 00.N- 0F.0- 06.0- 90. N0.0F 0m
00. F- 00.0 00.N 0N.0 06.6 00.N 00.0 06.0 00.0 0... F 00.6 06.0 00.0- 00. F 00.0 00. F- 00.3 6N
F0.0 0F.0 00.N 0 F0 0F.0 00.0 00.N 0N.0 06.N 00.0 0~.N 00.N 00.0- 0N.N 00. F- 00.N F». FF mm
00.0. 00.0 00.6 0N6 06.6 06.F 0 F0 00.0 00.6 00.0 00.6 00.0 06.0 00.0 00.N- 0N.0 0N. FF «N
0 F.F 00.N 00.0 0F.0 00.0 0N.N 00.N 00.0 00.N 00.6 00.0 06.6 0F. F- 05.0 00.0 00.0 0~.0F FN
0F.F 0~.0 00.N 06.0 00.6 00.0 00.0 00.N 00.0 00.0 00.6 00.0 06.0 00. F- 00.F 00.0. 00..- 00
0N. F 00.0 00.0 06.6 00.0 00.N 00.0 06.0 0~.N 00.0 00.0 06.0 00. F 9.0.. 06. F on. F 60.0 0F
0~.0 0N.F 00.0 00.0 00.N 00.0 06.0 00.6 00.0 00.0 00.6 06.6 00.N 0~.0 00. F- 0 F6 00. FF 0F
0F. F- 0F.0 0~.0 0N.0 06.0 00.N 00.0 00.6 00.6 00.6 00.0 00.0 00.0 00.0 00.0- 00.0- 0F. FF NF
0F.0- 00.0 0F.0 00.6 00.6 0F.6 00.0 00.0 00.N 00.6 0N.N 00.N 0N.0 00.F 06.0- 06.0- 00..- 0F
0 F.F 00.N 00.F 00.F 00.0 06.N 0N.0 0F.0 00.N 00.6 05F 00.0 0...F 06.0 00.0 00.0 50.0 0F
00.0 06.N 00.N 00.N 00.0 00.6 00.0 00.0 00.0 00.N 00.0 0~.N 00.6 00.6 00. F- 00.6 00.0 6F
0N.0 00.F 00.F 00.0 00.0 00.0 0F.6 00.0 00.N 00.0 00.0 00.6 06.N- 00.N 0F.0 0F.0. 60.0 0F
0N.0 00.F 05F 06.0 0F.0 00.0 06.0 06.0 06.N 0F.6 00.0 06.6 00.N- 00. F- 0F.F 06.0- F0.\- NF
00.0- 06. F 00.N 0F.6 06.0 0F.0 0~.0 00.0 00.0 0F.N 00.6 06.0 00.0 00.N- 00.6 06.0- 6... FF FF
0F.0. 00.N 06.6 0F.6 00.6 00.6 0F.6 00.N 00.0 00.0 06.0 00.N 00.6- 00.0 0F.0- 00.0 ~00 0F
00. 00.0 05.0 00.0 00.0 006 00.6 0~..F 00.0 06.0 00.0 00.0 00. F- 0~.6- 00.F 00. F- 00.NF 0
50. F 00.0 00.6 006 06.0 00.0 00.6 0N.0 00.0 00.0 00.6 00.0 0~.0- 0N.0 00.6- 060- 06.0F 0
«0.0 06.F 00.0 00.0 05.6 00.0 00.0 00.0 00.0 06.0 00.0 00.0 06.0 00.F 00.0 0F.0 0F.0F \-
66.0 05F 0F.6 00.N 06.0 00.N 06.N 0N6 00.0 06.0 00.0 0 F.N 00.0- 00.F 00.0. 00.0- 00.NF 0
0 F.F 00.F 00.N 006 00.6 06.N 00.6 00.0 00.0 0N.0 06.6 00.0 00.F 06.0 00.0. 06.F F0... 0
N~.0 0N. F 05.0 00.0 00.0 00.6 00.0 06.N 0~.N 00.0 00.0 00.0 0~.. F- 06.0. 00.N 00.0. 00.0 6
00.0 00.0 0N.0 0~.0 0F.6 06.0 0~.0 00.N 00.0 05.6 00.0 0F.N 00.6- 0N.0 0N6. 00.F 60.6F 0
0F. F- 00. F 0~.0 06.0 00.0 0~.N 00.0 0~.0 0~.N 00.0 00.0 00.6 05.0. 0~.. F- 0F.0- 00. F- 00.9 N
00.0 00.0 00.N 00.0 05.6 00.6 00.6 06.6 0N.0 06.6 0~.0 00.0 00.0- 0~.0. 00.0 06.0. 00.0 F
0.02 6P0”— 600—h- 6”..— 6“|— 6”..— 6“4 6W4 6"J 6M4 6W..— 6? mm 0X6“ ”EMF exp-hr 063n=§° 63¢

 

 

003001.002 . 000:0..0000 - 00:00< 09...... - 50.00505. 00:0..05 .00 0.00...

104

I”. . . vE.E. .I!

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.F 0N.0 00.0 00.0 0N.0 00.6 00.0 0F.0 006 00.F 0N.0 00.0 00.0 00.6- 000 05.N 0N.0F 0N
06. F- 00.N 00. F 00.0 0N.0 0F.N 00.0 05.0 00.N 00.0 00.0 00.0 00.0 00.0 06. F- 06.0 00.0F 6N
0N. T 05. F 00.0 00. F 0N.N 00.6 00.6 00. F 0F.0 0N.0 00.0 00.6 0N.N- 00. F- 00. F- 0F. F- N0.0F 0N
05.0 06. F 06.0 0N6 0 F .0 05. F 00.0 00.0 00.0 00.0 00.6 05.6 06.0 05.0. 00. F 0N.0 00. FF NN
0F.F 00.0 00.0 05.6 0N.0 05.F 00.0 00.0 00.F 00.0 0N6 00.6 00.0 0N.0 00.F 00.F- F5.0 FN
50. F 00.0 05.N 0F.0 00.0 0N6 0 F.0 00. F 05.0 00.0 006 0F.0 06.0 0F. F- 00. F 00.0- 00. FF 0N
N0.F- 06.6 06.0 00.0 00.N 00.N 0F.F 05.F 0F.0 006 00.0 05.0 00.F- 00.F 05.F- 00.F 50.6F 0F
F0. F- 00.N 00. F 0 F0 00.0 06.0 00.0 0F.0 0N6 0F.0 00.0 00.0 00.N 00.6- 05.0 00. F F6.0F 0F
N00- 0F.N 006 00.6 00.0 00.N 00.0 00.6 0N.0 06.6 00.0 06.0 00.0- 0N.0- 0N. F- 0F.0. 05.0F 5F
0N. F- 00. F 00.6 00.0 00.6 0N6 0F.0 06.0 00.0 00.0 00. F 0F.0 0F.0 00.0 0F.0. 00.6 0F.6F 0F
00.0 00.0 00.N 00. F 00. F 06.0 0N.0 0N6 00.0 00. F 00.6 00.0 06.0- 00. F 00. F- 0F.N 60.0F 0F
00.0 00. F 06.0 05. F 00.0 00.0 0F.N 00.6 00.N 00.0 05.0 0F.6 00.N 00.6 0F. F 00.0 06.0 6F
00.0- 0N. F 06. F 06.0 00.0 05.6 00.0 006 0N6 00.0 006 00.6 00. F 00.0 0F. F 05.0 06.NF 0F
00.0 00. F 00. F 00.N 05.6 06. F 0F.0 06.6 00.0 00.F 00.0 00.0 0F.F 0N.0. 0F.F 06.0- 00.0F NF
00 .0. 000 00.N 00.0 06.N 00.6 06.0 00.0 0 F .6 00.N 00.6 00.0 06.0- 06.0- 00.0 00. F N5.0 FF
00.0 00.0 00.0 00.0 06.0 05.6 05.6 06.0 00.0 00.0 00.0 0N.0 0N.0- 00. F 00.6- 00.0 05. FF 0F
0N.0- 00.0 00. F 00.0 00.0 00.0 00.6 00.0 05.6 00.0 00.N 006 00.0 00.6 00. F 00.0- 000 0
00.0- 00. F 00. F 00.6 00.N 00. F 00.0 00.0 0 F .0 00.N 05. F 00.0 06.F 06.0 00.0- 0F.0 05.3 0
00.0- 05. F 00.0 00. F 00.0 00.N 05.0 06.0 00.0 00.0 05.6 00.6 0N.0- 0F.0- 00.0- 00. F 50.2 5
0F. F 05.N 00.N 06.0 00.0 00.0 06.0 00.0 006 00.0 00.N 0F.0 06. F 00.0 06.0. 0F.0 05.0 0
05.0- 00.0 00.6 00. F 00.0 0N.0 06.0 0N.0 0N.0 0N.0 00.0 0N6 00.0- 00.N- 0 F. F 00.0 50.0 0
00.0- 060 00.0 00.F 0N.F 00.0 0F.0 0F.6 00.6 00.0 00.N 0F.0 00.N- 00.N 00.0- 00.6. 05.6F 6
05.0- 06.N 00.6 00.N 00.0 00.6 00.0 00.N 0N.0 06.0 00.0 0N.0 0F.F 00.0 00.F 00.F 00.0 0
00. F 0N. F 00.0 00.0 00. F 00.0 0N6 0F.N 06.N 06.N 00.N 00.0 06.0. 0F.0 00.0 06.0- 60.0 N
50.0 0F.0 00.N 00.0 00.0 0N.0 00.0 00.6 00.6 0F.0 00.6 00.6 0N.0- 00.0 0N.F 00.0 00.0 F
. 0 . .
.05. n... :0... a... a“. a. c“. :03 a". a”. s": 5.... PM. W...“ 0.... m... 23.8 s...

 

 

00.000”. .002 = 000023000 . 00:03 000...... - 50.000005. 0002.20 .06 0.00...

105

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N0.0- 06.F 00.0 06.N 06.0 00.N 0N.0 05.6 00.0 00.6 00.0. 05. F- 00.6- 0 F.N- 00.6- 00. F- 00.NF 0N
0 F .0. 00.0 00.6 00.0 00.0 06.6 0N.0 00. F 05.6 00.0 00. F 00.0 00.N 05. F 00.0 00. F 0F.5 6N
F0.0- 0 F .N 05.6 0N.0 00. F 00.N 00 .0 00.N 00.N 0N.N 05.6.. 00.N- 06.0 00.0- 00.0 00.6. 00.0F 0N
00.0- 00.0 00.0 00.0 00.F 00.0 00.0 0N.N 05.0 00.6 00. F- 00.0 00.0- 0N6 0N.0- 00.0 00.NF NN
00.0- 00.0 00.0 00.0 00.0 00.0 00.0 00.6 00.6 00.0 00.6- 0N.0 06.6 00.N- 00.0. 0F.0. 00.0F FN
00.0 00.0 06.0 0N6 00.0 00.6 0 F.0 0N6 00.0 06.F 05.0- 00.F 00. F- 0N.0- 05. F- 00.N- 60. FF 0N
50.F 00.0 00.N 0N.0 05.0 00.0 00.0 00.N 00.0 00.0 06.0 0N.0 00. F- 00.0- 00.0 05.0- 00.0F 0F
60.F 00.0 0N. F 00.0 00.0 00.6 06.N 05.6 06.0 0F.0 00.N 00.N 05.0- 0N.N 00.0- 00.0- 50.5 0F
00.0 05.0 00.N 00.N 00.6 0F.0 0N.0 00.N 06.0 00.6 0N.N 00.0 05.6 00.6- 00.F 06.N N5.0 5F
0F. F- 05.0 00.F 00.0 06.0 00.0 05.0 00.N 00.6 00.N 0F. F- 00. F- 00.6- 00.0- ON. F- 00.N- 00.0F 0F
50.0- 00.0 00.6 06.0 06.0 00.6 00.0 00.6 00.0 00.6 00.0- 00.F 00.N 00.0 00.N- 00. F 00.0 0F
F0.0 06. F 00.0 0F.N 00.N 06.0 00.0 00.0 00.0 0N6 06. F 06. F 00.6 05. F- 00.N 05.0 00.0 6F
N5.0 0N.N 0N.N 00.0 06.0 05.N 05.0 05.0 0F.0 0N.0 00.0 00.N 00.N 00.N 05.0- 06.0 0 F.0 0F
F0.F 00.0 00.0 00.N 00.0 05.6 0F.0 05.0 06.0 06.0 05.0 0N.0 00.N 00.0- 00.0 0N.0 0F.0 NF
F6.0- 00.0 06.6 05.N 06.N 0N.N 06.N 00.6 00.0 0N.0 00. F- 00. F- 0N.0- 00.0 OF. T 0F.N 06.0F FF
0N. F- 0F.N 00. F 00.0 00.0 00.6 00.0 00.0 00.6 06.0 00. F- 0N.F 0F.6. 0N6 00.N- 0F.0 F0.0F 0F
06. F 00.0 06.0 00.N 00.N 00.0 00.6 05.0 00.0 00.0 05. F 00. F 00.0 00.0 06.0 00.N- N60 0
NN.0- 00.0 00.0 0F.0 00.6 0F.0 00.0 00.N 00.0 06.0 00.0- 0N.0- 00.0- 00.N 00.0- 00.0- 00.0F 0
N5.0. 00.0 00.6 06.0 0N.0 00.0 00.0 05.N 00.6 06.N 06.0. 05.0 0F.0- 0F.o. 00.0. 00.0- 60. FF 5
50. F- 0F.N 00.F 00.0 00.0 00.6 00.0 00.N 00.6 0F.N 00.0- 06.N 00. F- 00.0. 00.N- 00.N- 00. FF 0
50. F- 00.0 00.0 00.6 05.6 00.N 0N.N 06.N 00.0 06.0 00.N- 0N. F 06. F 0N.0. 00.0- 00.0 N0.5 0
0N.0- 00.N 00.N 00.0 00.0 00.0 00. F 00.0 00.6 00.6 00.0- 00.N 056- 00.0- 00. F 00.6.. 05.0 6
0N.F 05.0 00.0 06.N 00.0 00.6 00.0 00.0 00.0 00.0 0F.0. 00.N 00.N 00.0 ON. F- 06.6 0N. FF 0
66.0 00.0 05.F 00.6 00.0 00.F 00.N 06.F 05.N 00.0 00.0 00.F 00.0- 06.0 00.0 00.0- 06. FF N
66.0. 06.N 06. F 00.0 05.6 06.0 0F.6 00.N 0N.0 00.N 05.6- 00.N 00. F 0N.0 00.0. 00. F 00 .6 F
. . . . . .
0.0.2 a”. cm. 0%.. cw.. c"... c". cm. cm. EF... «RM We“ We“ WM We.“ ”HF. 0.5030 :3.

 

 

0030012004. E 0.00:0...0000 - 0....00< 000...... - 50.5—.00.... 00:0..000 .F6 0.00...

106

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.0- 00.0 0N.0 05.N 0N.0 00.0 00.0 00.0 06.6 00.N 00.6 00.0 0N.0 00.N 00.N- 00. F- 00.5 0N
0 F. F- 05.0 0N.0 00.0 00.6 00.N 0N.0 0F.0 0N.0 00.0 00.6 00.0 00.N- 00.F 00.0. 06.N. 00.0F 6N
00.0- 00.0 00.N 06.0 00.0 06.0 00.N 00.0 00.N 00.0 05.N 00.N 00.0- 0N.N 06. F- 00.N 60. FF 0N
50.0- 00.0 00.0 06.0 00.N 00.6 05.6 0F.N 00.F 006 00.6 00.N 00.6- 00. F- 00.0. 00. F- 00. FF NN
F6.0- 00.N 00.0 06.6 05.0 00.0 00.0 00.N 06.N 00.0 006 00.6 00. F- 00.0 05.0. 05.N- 00.0F FN
0F. F 00.0 00.0 00.0 0F.6 00.F 0N.N 00.N 05.6 00.N 0N.N 00.F 00.N- 00.0 00.0- 00.0 05. FF 0N
N00. 0 F.F 0 F6 0N.F 00.0 00. F 00.0 00.6 05.N 00.0 0N.0 00.0 0F.0- 00.0 0F.6- 00.0- 0F.0F 0F
00.0 00.N 05.0 00.0 06.N 0N.0 00.0 0N.0 06.N 06.0 05.6 00.6 06.6- 00.F 00.0. 00.N- N0.0F 0F
00. F 00.0 00.6 00.6 00.N 0F.F 0F.0 00.0 00.0 00.0 00.6 0N.0 00.0- 00.0 00.0- 0N. F- F06F 5F
00.0- 00.0 05.0 00.6 00.N 0F.0 06.6 05.F 06.F 00.0 00.6 0N.0 06.N- 0F.0 00.0 00.N- 00. FF 0F
50.0- 05. F 0 F. F 00.0 00.0 00.0 00.6 06.0 00.0 00.0 00.6 00.0 00.6. 0N.N 05.6- 00.N- 0N. FF 0F
F0. F- 00. F 0N.0 00.F 00.0 00.0 05.0 0F.0 00.0 0F.F 06.0 00.0 05.0- 0F.0 06.0. 00. F. 50.NF 6F
0F. F- 00.0 00.0 00.6 05.N 00.6 00.0 00.N 00.N 00.0 00.0 00.6 06.0. 0F.N- 0F.N- 0F.N- 06.0F 0F
0F.0 0N6 00. F 0F.0 00.N 00.0 00.6 00.N 00.0 00.0 00.6 00.0 00. F- 0N. F- 06.6 00. F- 06.0 NF
0F. F 00.0 00.N 06.0 06.0 00.0 00.6 00.0 00.0 00.F 06.6 00.0 00.0 00.6 00.0- 00.F 60.0 FF
0F.0- 00.N 06.6 0F.6 00.6 00.6 0F.6 00.N 00.0 00.0 06.0 00.N 00.6- 00.0 00.0- 00.0 00.0F 0F
00.0 0F.N 00.6 00.0 00.0 05.0 06.0 00.0 0 F0 06.0 00.0 00.0 00.0- 00.0- 00.0- 00.6- N0.0F 0
06. F- 00.N 00.0 00.6 00.F 00.0 00.6 00.0 00.F 0 F.0 00.F 06.0 00.F 0N6- 00.0 05.0 60.NF 0
0F. F- 0N.0 0 F. F 00.F 00. F 0F.0 00.N 05. F 0N.F 0F.0 00.6 00.0 00.6- 05.0 00.0- 05.0 00.0F 5
00.0- 00. F 00.0 06.0 06.0 00.F 00.N 00.N 00.F 00.0 05.6 00.0 06.0 00.0 00.N- 0N.0- 5N. FF 0
0F.0 00.0 0N.0 06.6 00.0 00.N 0F.0 05.0 0F.0 00.0 00.0 05.0 00.F 00.0. 00.F 00.F 05.0 0
0F.0- 00.0 05.0 0F.0 06.0 00.6 00.6 0N.N 00.N 00.N 00.6 0F.0 00.N- 05.0 00.N 06.0 0N.NF 6
50.0- 05.0 06.N 00.0 00. F 00.0 00.0 00.0 00.F 00.0 00.6 00.N 00.N- 00.0. 0N.0 0N.0- 0F.0F 0
00.0- 00. F 00.0 0F.0 06.0 06.N 05.N 0N6 00.6 05.F 00.N 05. F 0N.0 05.0. 0F.0- 00.0. 6F.0 N
00.0- 000 0N.0 05.0 00.6 0N.N 006 0F.N 00.0 0N.F 00.6 0 F6 0N.F 00.0 00.F 00.F 00.5 F
4 ﬂ 0 0
.05. .0... ea... c”. c... s... a“. a... an. cm. s... s... 0......- R. 0.?me 2.3.8 5...

 

40.0000 .50... . .002. . 0.50.. 09...... . 50.028: 0002.20 0.. 0.00.

 

 

 

 

 

107

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0N. T 06. F 00.0 0N.N 06.0 00.N 0 F0 00.6 00.6 00.6 00.0- 05. F- 00.6- 05. F- 0 F6. 05. F- 0N.NF 0N
50.0- 00.0 00.6 00.0 00.0 0N6 00.6 00.F 05.0 00.0 00.F 00.0 0N.0 00.F 00. F- 0F.F- 00.0 6N
00.0 00.0 00.6 00.0 00.F 00.F 05.0 0F.6 00.N 05.N 05.6- 06.0- 06.0 0F.6- 00.N 00.0- 05.0F 0N
50.F 00.0 00.0 0F.6 00.0 05.6 06.0 00.F 0N6 05.0 00.6 05.F 00.6- 006 00.6- 0F.N 00.0 NN
00.0- 05.0 0F.N 0N.0 0 F0 00.0 06.6 06.N 05.6 00.0 00.6- 05. F- 05. F 00.0- 90 0N.N- F0.5F FN
00.0- 00.0 05.0 00.6 05.0 00.0 00.6 0F.0 0N.0 00.0 06. F- 05.F 00. F. 00.0- 06. F- 0F.N- F0.0F 0N
00.0- 00. F 0F.6 00.0 00. F 00.6 06.6 00.N 0N.0 066 00.N- 0N.N- 05.6- 00.0- 06.6 00.0- 06.N F 0F
50- 0N.F 06.0 0N.0 05.0 05.0 00.0 00.0 05.0 00.0 00.6- 00.6- 00.0 00.6- 00.0 00.0- 0N.0 0F
60.0 00.0 06.N 00.0 06.0 0F.6 00.6 00.0 00.0 06.0 00.0 00.N 00.0. 00.N 0N. F- 0N.N 00.6 5F
05.0- 0N.F 00.6 00.0 00.N 0N.N 00.6 0N.0 0F.N 00.N 00.N- 06. F- 06.0- 00.0- 00.N- 0N.N- N0.5F 0F
00.0 00.F 00. F 06.0 00.0 006 00.0 00.0 00.N 00.0 0N.0 00.0 06.0. 05. F 00. F- 00.F 06.0 0F
N5.0 00.F 06.6 0F.N 0F.N 0F.0 00.0 06.0 00.0 00.6 06.F 00.F 00.6 05. F- 00.N 05.0 66.0 6F
0F. F- 06.F 00.6 00.0 00.0 00.6 00.6 00.6 0N.0 00.0 0F.6- 00.0 0F.N- 00.6- 00.6. 0N.0. 0N.0F 0F
06.F 00.0 06.0 00.N 05.0 05.6 06.0 0F.F 06.0 0F.0 00.0 0N.0 00.N 00.0- 0F.6 00. F- 05.0 NF
00.0- 0N.N 06.6 00.0 00.0 05.0 00.6 06.6 0F.0 06.6 0N6- 06. F 00.6- 0F.6- 0N6- 00.0- 00. FF FF
00.0- 0N.N 00.N 00.0 00.6 00.0 00.0 006 0F.6 00.6 00.6- 05. F 06.0- 00.0- 0N. F- 06. F - F5.0 0F
NN.0- 00.0 00.6 00.N 00.N 00.0 06.0 06.0 00.0 00.6 06.0- 0N. T 0F.N- 00.0. 00. F- 00.0- F0.0F 0
00.0 00.0 06. F 05.6 06.0 00.0 00.0 0F.0 00.0 00.0 05.0 00.0 06.0- 20. 00.0- 0F.N 0F.0 0
66.0 00.N 05.N 00.N 00.0 00.0 05.N 0F.F 0N6 00.6 00. F- 06.0 05. F- 00. F- 05.0 0N.N- 00.0F 5
60. F- 00.N 05.N 00.N 00.0 00.0 00.0 0F.N 06.0 00.0 05.N- 00.N 00. F- 00.0- 00. F- 00.0- 00.NF 0
00.0- 00.0 0F.N 00.6 0F.0 00.0 0F.0 0N.0 00.0 0 F0 00.0 00. F- 00.F 00.0- 0F.N 00.0- 0F. FF 0
60.0 00.0 00.0 00.6 0N.0 0N.0 0N6 00.N 06.0 05.6 0N.N 00.0 00.N- 06.N 00. F- 00.F 5.2 6
50. F- 00.N 00.0 00.0 00.0 0N6 0 F0 0F.N 00.6 00.6 00. F- 00.0- 00.0- 00.0- 00.0- 05. F- F0.NF 0
0N.0 05.F 00.6 00.F 00.6 05.0 00.6 00.6 00.N 00.6 00.0 0N.0. 0N.0. 00.0. 0 F0. 00.F 00.0F N
00.0 00.0 00.0 06.0 00.0 00.0 00.N 0F.0 00.6 0F.0 0F.N- 05.F 0F.0 00.0 00.F 00.F 0N. FF F
. . . w . .
0.92 0". cm. 0%.. cm. cm. a"; cw. cm. EFJ «9.0.1 “an. PM. 00”.“ “>0”. ”M.“ 25030 :3.

 

 

0.3000 .52 ._ 002. . 0.000... 09.0... . 50.02.00... 0002.20 .0... 0.00.

108

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0N. F - - - - - - 00.0 0F. F 00.0 00.0 0F.0 - - 06.0 00.N 00.0- 0F.F 60.6 06 0N
00.0- - - 05.0 00.N 00.6 05.0 00.0 00.0 05.0 00.0 06.6 00.0. 00.N 00.0- 00. F. 00. F- 00.0- 60.0 N>> 6N
00.0 - - - - - - 00.F 00.0 00.0 0N.0 0N.0 - - 00.6 00.F 0F.0- 0N.F 06.0 06 0N
66.0 - - 00.0 0F.0 05.N 06.6 06.0 00.F 00.6 00.0 06.0 00. F 00.0 00.0- 05.N 0N. F- 06.N 05.0 00 NN
60. F - - 00.0 05.0 05.N 06.0 00.0 06.0 00. F 00.0 0N6 00.0 00.N 00.0 00.0- 0N6 05.0- 6N.0 00 FN
F6.0 0F.N 00.0 0F.6 00.6 00.0 06.0 00.6 0N.0 0F.0 05.0 0N.0 - - 0N.F 00.F 00.N 0N.0 00.0 Nm 0N
0N.0 - - - - - - 00.0 06.N 05.N 0F.0 00.6 - - 00.N 05.0 00.0. 00.F 06.0 06 0F
05.0 00.0 00.0 00.6 05.0 00.0 00.0 05.6 00.0 00.N 00.6 00.6 - - 0N.N 05.0 00.0 00.F 06.5 Nw 0F
60.F - - - - - - 00.0 00.N 0N6 00.0 05.0 - - 05.0 06.N 06.0- 05.N 60.6 06 5F
0F. F 00.F 05.N 006 0N.0 0F.N 00.6 00.N 00.0 00. F 00.0 0N6 - - 00.0 00.N 00. F- 00.N 00.6 Fm 0F
50 - - - - - - 00.0 00.F 00.6 00.0 00.0 - - 0F.6 00.F 05.0- 05.F 00.6 06 0 F
00.0 00.0 00.6 00.N 00.N 06.0 0N.N 00.0 00. F 00.6 00.6 0F.0 - - 00.0 00.0- 00.0 0F.0 06.0 Nm 6F
0N. F 00. F 00.0 0F.6 0F.0 00.N 00.0 00.6 00.0 0N6 06.6 00.F - - 0N.F 00.0 00. F 00.0- 05.0 Fm 0F
50. F- - - - - - - 00.F 00.F 00.0 06.0 05.0 - - 00.0. 00.N- 00.0 06. F- 60.0 06 NF
0N. F - - 00.0 05. F 00.0 00.6 00.F 06.N 00.N 05. F 0N.0 00.0 00.6 00. F- 05.F 05. F- 0F.N 0F.6 00 FF
50.0 - - - - - - 00.F 05.F 00.0 0F.0 00.0 - - 00.6 00.N 00.0- 00.N 5N.0 06 0F
56.0 0N.0 0N6 06.6 00.N 06. F 00.N 00.N 00.0 06.0 00.6 00.6 - - 00.0- 05.F 05. F- 0 F0 0N.0 Nm 0
0N.0- - - 00.0 00.0 0N.0 00.0 06.6 06.0 00.N 00.0 00.0 06.N- 06. F 00.0- 00. F- 00.N- 00. F- 00.0 N>> 0
0N.0 00.0 00.N 00.0 0F.N 00.6 00.0 0F.0 00.6 0N.0 00.0 00.0 - - 06.F 00.0- 06.0- 06.0 F0.5 N0 5
0F.F - - 00.0 00.0 00.0 00.0 0N.0 05.0 00.F 05.6 00.0 00. F 0N.0- 00.0 0N.0 00.0 06.N 66.0 00 0
56.0 - - 00.N 0F. F 05.6 00.0 06.0 0N.0 00. F 06.0 00.6 00.6 0N6 0F.0. 00.N 0 F0. 00. F 0N.0 N>> 0
0N.0- . - - - - - 0 F.F 00.F 00.6 00.0 05.N - - 00. F- 00.0 05.0 05.0. 60.0 06 6
0F.F - - 00.0 00. F 00.0 00.0 0F.0 00.0 00.0 00.N 00.0 06.6 00.6 0F.0 05.0 00.F 00.0 06.N N>> 0
05.0 - - 00.0 0N.0 06.0 06.0 06.0 00.0 00.0 06.0 0N.0 00.6 00.6 0N6- 006 05. F- 00.6 56.6 N>> N
0N. F - - - - - - 00. F 0F.0 06.0 0N.0 06.0 - - 05.6 00.6 0F.6- 0F6 00.0 06 F
202 690—”.— 6Oo_n— 6MJ 6WJ 6W-— 6“..— 6WJ 6"4 6M.— 6”..- 6".— ”6“ “6” “>6“ WM.— u>6hH exp—M. 063lmao 0gb. 63¢

 

 

 

 

0.3001050... 0.0.9.0230 - 50.005002 0023.20 .66 0.00...

109

Appendix E - "Known Solution" Mechanism Case Study - Raw Data

110

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0N.0- 50.0 00.0 56.6 00.0 00.0 50.0 50.0 0F.0 0F.0 00.0 50.0 00.0 00.6 F0.0 00.0 00.0 06. F- 0N.06 mm 0N
00.0- 00. F 06.6 00.6 N00 60.0 00.0F 6N.0 00.6 00.5 00.0 00.0 06.0. 50.6 N0.0- 0F.0 N60. 05. F- 00. FN mm 6N
55.0- 06.6 50.0 00.6 05.6 NF.0 00.2 00.0 00.0 0N.5 00.0 66.0 00.6- 06.6 00.0- 50.0 00.0- 00.0- 00.6F mm 0N
N0.0- 00.0 :0 06.0 F0.0 00.5 0F.0 00.0 0N.6 00.5 00.0 50.0 5F.6. F6.6 00.0- 00. F- 50. F- 6N. F- 00.0N mm NN
50.0- F0. F 0 F .0 06.0 60.0 00.6 6F.0 6N.0 0F.6 55.0 00.0 00.6 60.0- 0N.0 50.N- 50.0 00.6- 50.N 00.00 mm FN
00.0- 00.0 00.0 00.0 60.5 00.0 0F.0 05.6 0N.0 0N.6 N00 No.0 00.0- 0N.6 00.0- N0.0 5N. F- 06.N- F0.5F mm 0N
00. F- 00.0 00.N 00.5 00.0 00.0 00.0 00.5 50. F 06.0 00.0 50.5 60.0- N0. F- F00- 00.6 F00- 05. F- 0N.N0 mm 0F
0N.0 00.0 5.0 66.0 06.6 05.5 00.5 00.0 60.0 N00 N60 60.0 50.0 65.6 5.0- 5 F.F N0.0- 00.0 00.00 am 0F
5N.0- 00.0 00.0 00.0 0F.0 5 F.5 00.0 F6.0 00.F N00 0N.0 05.0 00.0- 50.6 60.N- 00. F- N0.0- 00. F- 00.0F mm 5F
66.0- 000 00.0 00.5 00.0 N00 00.0 50.0 NN.0 66.0 00.5 00.N 50.0- 00.6 F6.0- 00.N 60.N- 06.0 0F.00 N>> 0F
60.0- 000 OF. F 00.0 00.N 05.0 00.5 F00 N5.0 00.N N60 60.N 60.0- 00.F 00.6- F6.0 0N.0- 06.N NN.00 mm 0F
N0. F- N00 00.0 N0.6 56. F 00.0 00.0 50.0 00.0 F0.0 50.0 00.0 00.0- 60.0 00.N- F0.0 00. F- 00.0- F000 mm 6F
00.0- N6.5 00.0 05.5 00.0F NF.0 00.5 05.0 0N.0 50.0 F00 00.6 N0.0 6F.0 N0.5- 60.0 05.0- N00- 00.0N .25 0F
00.0 N6.F 05.0 05.0 NN.N N5.5 06.6 5.0 00.N 6F.0 00.0 0F.0 00.0- 00.F 05.0- 5.6 00. F- 06.N- 05.NN H>> NF
N60- 00.6 N6.5 NN.6 0F.5 00.0 00.0 00.0 06.0 0F.6 00.6 00.0 0F.6- 06.6 05.0- 0N.0- 0F.N- 00.F 00.0N mm FF
05.0- 60.6 50.0 0F.0 05.6 00.5 00.0 N60 00.0 N00 50.0 50.0 00.6- 00.0 50.0- 65.0 00.0- 60.0- F0.NN mm 0F
00.0 06.0 00.F N00 N5.0 00.0 00.0 00.F 00.0 0N.0 00.5 00.0 5F.0- 00.6- 50. F- N F.6 50.0- 00.0 0N.N0 Nm 0
0F.F N0.0 0N6 00.0 60.6 00.0 00.6 00.0 00.N N0.5 0N.0 00.0 50.0- 06.0 0 F0. 00.0 06.N 00.N- 0F.0N H; 0
0F.0 60.0 5.0 0N.5 0N.0 FN.5 00.0 00.0 00.0 00.0 60.5 00.0 00.6- 00.0F 00.6- 00.0- 00.6 00. F- 00.66 Hm 5
0N. T 06.6 00.F 6N.0 00.5 00.5 00.0 5N.0 00.0 60.6 00.0 60.0 00.0- N0.0- 00.0- 5F.6 0F. F- 50.0 0N.N6 Nm 0
00.0- 00.6 N0.5 00.0 60.0 00.0 5.6 :0 00.5 00.0 50.0 05.0 00.6- 006 00.0- 050 N00. 000- 00.0N mm 0
00.0- 00.F 00.0 0N.0 0N.0 0N.0 06.N 0F.5 05.0 50.0 00.0 00.0 06.F 00.F 0F.0- 06.6 50. F- 00.F- 0F.00 H; 6
NF.0- 0N.0 50.0 00.0 F0.N N0.0 05.0 00.0 5.0 56.F 00.5 60.5 00.0 00.0 00.F 66.F N0.0- F0.0 60.60 mm 0
06.0- 50.0 NN.6 00.0 55.0 06.N 0F.6 5.6 00.0 00.N No.6 00.0 NF.6- N00 0F.0- 00.N 0N. F- 05.0 NF.0N H; N
00.0- 30 0N.0 00.6 00.0 00.0 05.6 50.0 0N.5 06.5 00.0 0N.0 60.0- FF.6- 0F.0- 50.6 NF.0- 00.6- 000 H; F
235 CFO—”d COO—n. C”:— CNJ CWJ CWJ C"...- C".— CMI- CM.._ 9.".— bwﬁ «Xv—M ”C“ “X...“ ”a.“ ”bi” ESE—Do EH 33¢

 

 

.338”. § .52 2552.8 .2320 . 5228: 223.8 535... .3 «33

111

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

66.0. 00.0 00.N 50.5 65.6 00.5 50.0 00.6 5 F.5 06.0 66.6 N0.N 00.0- 00.F N0.0- 00.0 00.N- 0N.0 00.60 00 0N
00 .0- 0N.0 00.0 00.0 50.5 00.0 50.5 00.0 60.0 05.0 0 F0 05.0 56.0- 5N.6 F0.0- 0 F0 60. F- 05.0 06.00 00 6N
F6.0- 0N.0 00.0 0N.0 00.5 0F.0 00.0 5F.5 00.0 F0.F 00.0 50.5 00.0- 50.6 00.0- 00. F- 00.0- 00.0 05.0N N>> 0N
06.0. 00.0 00.0 05.0 00.0 50.0 N5.N 50.5 NN.N 00.0 00.0 06.0 F0.0- 06.0 0F.0- 50.6 00.6- 000 0N.0N N0 NN
00 .0- NN.N 60.6 56.5 05.5 05.0 06.0 05.5 0N.0 F0.F 50.0 00.0 6F.5- 50.6 06. F- 00.0 60. F- 00. F F0.5N N>> FN
00.0 60.0 06.0 00.0 00.6 0N.0 60.6 00.0 60.6 00.0 00.0 06.0 06. F 0N.5- N0.0- 50.6 N0. F- 50.0- 00.6N F>> 0N
66.0- 05.0 F0.F 0N.0 F0.F 60.0 00.6 N0.N F0.F 0F.0 0F.0 66.6 00.5 00.0F 00.0- 50.0 00.0- 5N.6- 60.00 F>> 0F
60.0- 000 06.0 00.0 00.0 65.6 0F.0 00.0F 50.5 00.0 00.0 60.6 06. F- 00.0- 50.0- 000 F0.0- 00.F 00.00 06 0F
05.0 60.6 50.5 05.0 F0.F 00.0 0F.0 00.0 F0.F 00.0 00.0 00.6 6N.N 0N.F 00.6- 500 0N.0 50.0- 00.00 F>> 5F
F0. F- 00.0 00.6 0N.0 0 F0 00.0 00.0 50.0 05.0 50.5 00.0 06.0 00.6- 5N.6 00.5- 50.0 6N.N- 0N.0 05.0N 00 0F
66.0- 50.0 50.0 55.0 50.5 00.0 06.0 00.0 0 F .0 00.0 0N.0 05.0 N0.0- 50.6 F0.0 05.0- N0. F 00.N 0N.5 F 00 0F
00.0- 550 F0. F 05.0 00.0 66.6 60.6 00.0 00.0 00.0 06.0 00.6 00.6. 00.N- 6F.5- 0N.0 F0.0- 05.0- 0N.6N F>> 6F
0N.0- 00.0 05.0 00.0 00.0 50.0 00.5 00.0 50.5 N0.N 00.0 5N.5 F0.0- 00.0 6N.N- 00.0- 60. F- 00.5 00.06 00 0F
50.0- 50.5 00.0 N6.N 55.0 5N.5 0N.0 50.5 05.5 5N.5 00.5 00.0 00.N- 50.6- F0.0- 56.6 N0.F- 00.0- 06.NN 06 NF
0N.F F0.F 0F.0 00.0 06.0 05.5 00.0 06.0 00.0 06.0 00.0 00.0 5.0 50.0 00.0- 55.0 6N.N- 00.0 0N.60 N0 FF
0F.0 50.0 00.5 N6.N 0F.0 00.0 5F.5 00.5 50.0 05.5 N0.N 00.0 60. F- 00.5 56.0 0N.F 0N.0- 0N.0 00.0N N0 0F
F6.0- 00.5 N6.N 00.0 00.0F 00.6 0F.0 00.0 00.6 60.6 0F.0 0F.0 00.6 00.5- 00.5- 05.0 00.6- 00.0- 06.6N F>> 0
05.0- 00.0 0N.0 00.5 5N.5 0N.0 00.0 06.0 50.0 00.0 00.0F 00.5 6F.5- 50.6 00.0- 00.0- 00.0- 06.0- 00.0N 0m 0
00.0- 00.0 00.0 00.0 6N6 00.0 50.5 50.0 06.0 00.0 60.6 06.0 00.5- 50.0 60. F- 05.0 56.0- 05.N 0F.0N N>> 5
66.0- 600 60.6 00.0 00 .0 6F.6 N F .N 56.5 00.0 0 F .0 00.0 50.0 00.5 50.6- 00.0- 50.6 F0.0- 0F. F- 00.00 06 0
60. F- 0N.0 60.0 00.0 60.0 00.0 00.5 66.6 50.0 0F.0 50.5 50.5 00.0 00.0 0N6- 05.N N0.0- 5 F0 00.00 N0 0
0N.0- 0N.0 00.0 00.0 5N.5 5 F.5 5N.5 06.0 0 F0 50.5 00.0 00.0 00.0 00.F 6F.5- 00.0 F0.0- 00.N- 00.0F F>> 6
0F.0 00.0F 0F.0 NF.N N0.N 60.6 N6.N N0.N F0.F 00.0F 0N.0 5 F.5 06. F- 00.0- 00.5- 06.F 00.N- 00.F N0. F0 F>> 0
N F.0- 0N.0 00.6 00.0 N0.N 66.6 0N.0 00.0 00.0 00.6 06.0 50.5 N00 50.0- 6F.5- 55.0 N0.F 00.F F0.0N F>> N
60. F- 50.0 60.6 F5.F 00.0 00.0 00.0F 50.5 00.0 00.0 50.0 05.0 6N.N- 00.0 F0.0- 50.6 N0. F- 0F.0- 00.0N 00 F
0.05 Crud C0094 Bun. C".— C”.— C“.— CWJ C".— CMJ EMA C".— p“ WNW ”CM “X...“ ”but «Xv—M Utah—DO EF :3“

 

 

.338”. as :32 25528 32.25 . 3:23: $2.28 535... .8 2......»

112

[1]

[2]

[3]

[4]

[5]

{6]

[7]

[8]

[9]

References

Ord-Hume, A. W. (1977). Perpetual Motion: The History of an Obsession.
Adventures Unlimited Press.

Forbes, R. J. (1958). Man the Maker— A History of Technology and
Engineering. Abelard-Schuman Publishers.

Singer, C. J. (1958). A History of Technology — The Industrial Revolution.
Oxford University Press.

Hrones, J.A., and Nelson G.L. (1951). Analysis of the Four-Bar Linkage.
MIT Press and Wiley, New York.

Freudenstein, F. (1954). "An Analytical Approach to the Design of Four-
Link Mechanisms”, Transactions of the ASME, vol. 76, pp. 483-492.

Mallik, A.K., Ghosh, A., and Dittrich, G. (1994). Kinematic Analysis and
Synthesis of Mechanisms. CRC Press.

Erdman, AG. (1995). "Computer-Aided Mechanism Design: Now and the
Future", Transactions of the ASME Journal of Mechanism Design Special
50th Anniversary Issue, vol. 1178, pp. 93-100.

Erdman, A., Sandor, G., and Kota, S. (2001 ). Mechanism Design —
Analysis and Synthesis. vol. 1, 4th edition, Prentice Hall.

Mermertas, and Vahit. (2004). “Optimal Design of Manipulator With Four-
Bar Mechanism”, Mechanism and Machine Theory, vol. 39, issue 5, pp
545-554.

113

[10]

[11]

[1 2]

[13]

[14]

[15]

[16]

[17]

“Ti—EVEN...“

Awtar, S., and Gupta, V. (1997). “Synthesis of Four-Bar Mechanisms
using Pattern Matching Approach and GA search”. Indian Institute of
Technology, Kanpur (Class Project.)
http://www.mit.edulpeople/shorya/BTech/GA.htm

Vasiliu, A., and Yannou, B. (2001). “Dimensional Synthesis of Planar
Mechanisms Using Neural Networks: Application to Path Generator
Linkages", Mechanism and Machine Theory, vol. 36, issue 2, pp 299-310.

Krishnamurty, S., and Turcic, D.A. “Optimal Synthesis of Mechanisms
Using Nonlinear Goal Programming Techniques”, Mechanism and
Machine Theory, vol. 27, issue 5, pp 599-612.

Holland, J. H. (1962). "Outline for a Logical Theory of Adaptive Systems",
Journal of the ACM, vol. 9, issue 3, pp. 297-314.

Back, T., Fogel, D.B., and Michalewicz T. (2000). Evolutionary
Computation 1 - Basic Algorithms and Operators. Institute of Physics
Publishing Ltd.

Fang, W. E. (1994). "Simultaneous Type and Dimensional Synthesis of
Mechanisms by Genetic Algorithms", American Society of Mechanical
Engineers, Design Engineering Division, 23rd Biennial Mechanisms
Conference, vol. 70, pp. 35-41.

Connor, A.M., Douglas, 8.8., and Gilmartin, M.J. (1995). "Kinematic
Synthesis of Path Generating Mechanisms Using Genetic Algorithms”,
Applications of Artiﬁcial Intelligence in Engineering, pp. 237-244.

Roston, G., and Sturges, R. (1996). "Genetic Algorithm Synthesis of Four-
Bar Mechanisms", Artiﬁcial Intelligence for Engineering Design, Analysis,
and Manufacturing, vol. 10, issue 5, pp. 371-390.

114

[18]

[19]

[20]

[21]

[22]

[2 3]

[24]

[25]

Tachiya, H., and Tsujimura, Y. (1996). ”Synthesis of Planar Multilink
Mechanisms Based on Genetic Algorithms", JSME International Journal,
Series C, vol. 39, issue 4, pp. 828-836.

Kunjur, A., and Krishnamurty, S. (1997). "Genetic Algorithms in
Mechanism Synthesis", Journal of Applied Mechanism and Robotics, vol.
4, issue 2, pp. 18-24.

Koladiya, D., Shiakolas, P. S., and Kebrle J. (2003). “Evolutionary Based
Optimal Synthesis of Four-Bar Mechanisms", Proceedings of the ASME
Dynamic Systems and Control Division - 2003, vol. 72, issue 1, pp. 539-
544.

Cabrera, J.A., Simon, A., and Prado, M. (2002). "Optimal Synthesis of
Mechanisms With Genetic Algorithms", Mechanism and Machine Theory,
vol. 37, issue 10, pp. 1165-1177.

Liu, Yi. (2004) Automated Type and Dimensional Synthesis of Planar
Mechanisms Using Numeric Optimization Wrth Genetic Algorithms. PhD
Dissertation, University of Waterloo.

Liu, Y., and McPhee, J. (2005). "Automated Type Synthesis of Planar
Mechanisms Using Numeric Optimization With Genetic Algorithms",

Journal of Mechanical Design, Transactions of the ASME, vol. 127, issue
5. pp. 910-916.

Lin, S.-C., Punch, W., and Goodman, E. (1994). "Coarse-Grain Parallel
Genetic Algorithms: Categorization and New Approach", Proceedings of
the Sixth IEEE Symposium on Parallel and Distributed Processing. pp.
28—37.

Red Cedar Technology (2005). HEEDS User’s Manual - Version 5.0.
www.redcedartech.com

115

[26]

[2 7]

[28]

[29]

Gustafson, S. (2006). "The Speciating Island Model: An Alternative
Parallel Evolutionary Algorithm", Journal of Parallel and Distributed
Computing, vol. 66, issue 8, pp. 1025-1036.

Chen, EC. (2005). "Analysis and Veriﬁcation of a Watt I Six-Bar
Furniture Hinge Mechanism", Proceedings of the l MECH E - Part C -
Journal of Mechanical Engineering Science, vol. 219, issue 10, pp. 1107-
1 1 17.

Shieh, W.-B., Tsai, L.-W., and Azarm, S. (1997). "Design and
Optimization of a One-Degree-of-Freedom Six-Bar Leg Mechanism for a
Walking Machine", Journal of Robotic Systems, vol. 14, issue 12, pp.
871-880.

Yu, N., Erdman, A. G., and Byers, B. P. (2002). "LINCAGES 2000
Latest Developments and Case Study", Proceedings of the ASME Design
Engineering Technical Conference (Special Session on Computer Aided
Linkage Design), vol. 5B, pp. 1421-1425.

116