30:3..— uh : ". 2 I] 5}”; (:1?! 572': «L. ”7"". .'4 F‘I 3% fl 3.94“ k.“ ML"i gin!“ ”Q 't ~ .39.:va 0 u}, ‘rfis' ' (II: 2 4‘5 79* . It): I ilfivx'tr J1" I I \r I! .I ,:_‘{ .llf.‘ ('I " ' ’Jnr'wrk 'II" 'I 3" ~. I'm ' 9'“ r1 < ‘ {I I. fix" 5': 5'33“ fi'fi' -.4_,:. ‘-' ($3.95; “‘0‘: C): "I * n " Anv- "Iii-.I‘f'fv‘igg w}. 1 ,- I44 -4. ' "i' ‘Ui'h '1 :I_'{II'-"2{:§Ja:p.,’ . . ‘2'. .2 4,?! 4. .'I:|I"".,-7-I""14'4'JJ‘fl'I 4.51.1. ' , «"44! *“ 144.4 Hg .4: 4 """‘.' , I'4 gII ”wry-Ira"5.,iLII’4’ .' " , ' " '1‘“.II "' 'J' ,“":"'I" ' ' .4 I 4... III . .. ; I. I . 4+, 1521“ ' 4 :: -"I:,!I;1“’II.';:" 'I A“ 1,. I. . '\'.::’I“,|‘ . ' “5'2: \\v 135;? Egg” fia'IT'K. lag/E4, I $4” "(12' IN. Q‘}~1,I TREE? 2:1»)? 15 3f": 2%;:: 5:5" 1:01;“- .'"' .V '3'} 1&1}... fi'ifi‘fiv Ski." f4. \{2 rm. E131.“ '4" 33‘ '9' N "Z“, '{i~ 8""‘ \' ’JILI n 4 WW. fig?) 5’5 ' “1. . 31%?!“ {£54. .P, 'u: I). was I _. 2,“: '\ 2 a} _ A. ~52.- ‘ .7. 'Q J 7‘4: .',-,','-:: Jpfi-D; g'rr‘zi 4 “1-" ‘ a.“ #3:." a 375.2--- Jv- ..J;‘: :v'f— — ii. -m?‘ . 3+1; .‘ .1 - JR ' y ‘n {5‘ C ‘ _‘ 4“ :4' . . .41. 'S'II I}? 915“: a, 4 4; 44-- r'! ' I 6'36, {J4 t. "CHI. It, ." I, II "5' SE"? ,1?th ”I“ '14:? . ‘4’"... . '9' ' " I' I ' "w 'I' 3"" r7"'I "'J """"" J \ ."IT'M "Iqt‘I'ri' "Vi", ‘ " T £0.14" "'4'5 ",5!“ l'fi"4.I,' Va"! 'd' "" " "'5 "' 3 "i " "Ik " "I. "' ' " ' ' ‘ u 't I ”:91 l '| 'I'. "' "' . 2"“, '4 ‘L "I :l, ‘ 11‘4" ' a" 54;- |" :J E" . H" ‘5! .r l “|")\ U‘QI, {4&1} {I :1"): :3' I: ' .t‘l 'II‘ 4 4:,445 4 " u" 4‘? #le ~ I.” -. a f, )1. ")'.'l" 414.] .. I I“); 41"‘4‘0 ' .. ."|' 'I “g 'n- 334‘. "n ' {c.‘k I‘I "" 1,45 .' " " ""‘Hll ..I""’ ‘4' "1",, . "'H "Y” :J. v"; 4 "MIN ' ' Hm“ -|'p'I \’ "'I..'.'.:' ' " ' {Y'IM '..~ “4' ' ‘-,"' ‘4 ‘.- J'I‘ 4,344" ‘ '4'. r A I 1‘. ‘ -" . ,. 'I.‘I'; I , I .’ ." H "' '4 ' | 'J 'I "(.'m 4 H." "" "'r" . "'(HIH:1""‘ ' I'":"" "f; ': " I: ' In I I ~‘ 49.. .4 ‘."II'.;1I'jm_' ~".f TI", ~ ‘ "‘ 14'. I’m 31:4 fit-I‘m‘ ' ' ""‘ '4 .- I -. . . i ' 4 I"': I "km"? 4 ‘1'"(JJ'L "4' ' " I""' " " '1 w "' I 4' '-':"4'}’$;"'u' ""1""'"""':.‘ 4"" ' 4" '13,"?! 2"“ V'" dr "':'?"v' 53"" r" . ‘ '4'. I .' I' I, : "II III I '4.' . .]~ ‘ '. II .'I"m'_' .I "I t; l 17' . ' f" L. ' "Tk"':‘." ""I ‘7]; .4“ .. 3w 14%;? 'I" 1 .""':I""' 'J 'dl' ' I ' "h '4'"JI"' 4' J ;"".I 3': "H I I 'Il'l "'HJ ' ' 'i' ' I ".'."' " '" ‘9'.""3'5 U"" ’ ”1": " - '8'. I 'fi'w’tfi'".}:l - I \ tg'w 'p' .H' .‘i' ‘0'”: J :4 ‘I I] ‘ 5-1 7.14,...” I", ,1”. 1,42,: {'1' .u'f: " |.'.'f""n,'-'ll ..;“ I'Ji'h 7. l . g, “45v l,- .1. 21-" .‘. 4,5: 7",. .4 "5:1" 4:, .it”.z'1;.‘.ri"','.fi,!.;:fi‘j,lhurt-5;. 3;,- , £~ I 4 ,1'. * ‘3 H1- 3.1. 4'51. ‘5‘, “.1 f' ‘u :', LIL-l 3” .3!" .» f" 1 '. M" . -;’ 31".:0 '3 I. 4,». ’11-: """""'I'3"J'I;" 4.5;.g1I:-';n,: ,2;:;.§,;.;.,I‘4n.¢{1,-t- ["4 , ., ‘ , .... _| HI I ' , '.4 ‘ ”‘44!“ . h 1| ”I. “‘ J" 0"..I-Ilulu. .. - f ‘-' . ..‘ l"‘ L. ‘l I. . I '1“ III- Inn", .0. .“j . , l . (1- -. h‘ g. I‘ ‘4' '1 ,"Ju‘s “A '-.".I'{-"' III, 0%. 4%.] l"' 7! ; ' I 'J "'1"1!.' ' ' V" ' '4 .‘ J H '. (II '11." h "f: 13'" "}"" I, ' "\'""""' ‘f' f .6 """",,,')!'.i' ;\ | I I‘lI'U,” l I :I" t '| . ""'"‘E§g('£'§'%|{'i" 4h} T‘R‘" fitsssiflllahé ' ' H H 1 , " _ “I“, ' , -‘u"'_"I J: M4 .\ 'n'": ' 4 ":‘f: I. |.‘4 44' :I: . . J... ' %!,l._l,q"‘él'1‘ ”‘1‘: 'f.|"4"' an? '1‘;"é?:44',§3‘:'2:"-£ I" “01"” I ' gm,» ”.'Iéji‘fig. “45'.- ‘II: ‘ --I§,IQ’.:5‘I""I' 4,;_; . 4.1%?" ”'"I' 'I‘N -I 4.19% —. I .. ,..~".;r" ,.".."".‘ 'I'I .. 4 .4 'J‘r '-' ,Ffi' {5,4133% .} {aghfi 't .4 I ' 5 ' "'15; I . .' I‘ I 4 . I‘. . _ I4. '9‘“ I {NH} '."';*"§" .' .343 “.'iij' '.' J 'V' ' 'I'I'k" .‘ 'I "'I‘ " ,1, I “til. "'11: i ""'J,:.' ,4 ‘I'hl't' "" ., ' ":":‘:." ;"¢ .4... ‘l .‘ '1' ‘ ' "'4' 1| TI ~=t {_ ._4-_ A---) - fi". - .- -- ‘T‘. : 3; H .I. I. ... ...,,;I4,I,24.444, .. '- V‘ ‘7 :ww In {éfij‘i .. . . ' _, ‘ ’F'z:"' “‘7 ,' ‘ c‘ m "1"“ _4 II, . 44:. "I13,” 4 ”WWW I. “If" .u, - I'- 'I .II , H4 III-y .! . - '4 J .. ‘ . (NI 2”,}! .'- ' ,'.‘l ':"V'I4I.'4l, .I"I1.“~, . . I . . .I. . -. n ‘ .' _ .‘I. L, 'k': ll,'-.| . ‘I‘".4""4".'|'""V ".LfJ-"J': NH“ "4'!" rm. ' ”g "g" Ill. $5} " 'é-iL'. 1 " I ,J a." ”I "I [0:1, ['I' I. “I 'I'y,. “ ’..' 'L ,IK‘ ‘3': .. III""""(', I ~. '. .. 5, 4- II .4 W’ .t:« L ‘ ILL" '- II '.'.-‘« I ‘~-’ I' ' - "‘II‘IE-I‘W w...” ".. I‘. ‘ W m' ‘1. .Iu'. 4. '.I — I: E Z A; 4;“ ‘1’ 5 hi muffin 523,126 E c"‘\ University f V - _ yarn-gut.) i3 This is to certify that the thesis entitled An Analysis of the Energy Conservation Potential of Variable Work Hours and Alternative Transit Policies for the Urban Work Trip presented by James Mitchell Witkowski has been accepted towards fulfillment of the requirements for Ph.D. degreein Civil Engineering 41%;. ”fly/m? Datewflfl— % E $454.,— 0-7639 OVERDUE FINES: 25¢ per day per item RETURNING LIBRARY MATERIALS: Place in book return to runove charge from circulation records AN ANALYSIS OF THE ENERGY CONSERVATION POTENTIAL OF VARIABLE WORK HOURS AND ALTERNATIVE TRANSIT POLICIES FOR THE URBAN WORK TRIP BY James Mitchell Witkowski A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements ' for the degree of DOCTOR OF PHILOSOPHY Department of Civil and Sanitary Engineering 1980 ABSTRACT AN ANALYSIS OF THE ENERGY CONSERVATION POTENTIAL OF VARIABLE WORK HOURS AND ALTERNATIVE TRANSIT POLICIES FOR THE URBAN WORK TRIP BY James Mitchell Witkowski The primary hypothesis tested by this research was that a reduction in transportation fuel consumption for the urban work trip could be real- ized through the implementation of a staggered or flexible work hour program. It was also hypothesized that a variable work hour program could be coordinated with the scheduling of a bus transit system to fur- ther improve the savings in transportation fuel consumption. The spatial organization of a hypothetical urban area was generated using data from the literature and a computer simulation program designed to distribute population and employment activities throughout the urban area. With additional data describing the highway and transit network, and the temporal distribution of work travel, the computer program also generated the work trip travel pattern for the urban area and computed the transportation fuel requirements for automobile work trips and the daily transit service. A base case was generated and used as the basis for comparison of the alternative policies. Several alternative temporal distributions of work travel were used to simulate the effect of variable work hour programs. Tests were de- signed to determine the influence of the magnitude and location of the work force participating in the variable work hour programs on the re- duction in fuel consumption. Experiments were also designed to test the potential for reducing fuel consumption through modifications in the scheduling of transit vehicles during the peak travel period. Other James Mitchell Witkowski experiments were designed to test the combined effectiveness of variable work hour and transit scheduling policies. An evaluation was also made to determine whether or not the prevailing fuel environment could influence the policy effectiveness. The simulation results indicated a high potential for staggered and flexible work hour programs to reduce automobile work trip gasoline consumption. The effectiveness of the variable work hour policies was shown to be influenced by both the number of participants in the program and the dispersion of the participants throughout the urban area. The reduction in fuel consumption increased with the number of participating work travelers. The reduction also increased as the locations of the participating employment centers became more dispersed throughout the urban area. The variable work hour programs also showed a strong nega- tive influence on work trip bus ridership. The experiments with transit scheduling policies indicated that it would be difficult to reduce work trip fuel consumption through in- creases in the frequency of service of bus transit. However, this result may be biased by the overall structure of the hypotehtical urban area and the base case used in the analysis. The combined variable work hour and transit scheduling policies showed no further reduction in fuel con- sumption beyond that achieved by the variable work hour policies. The experiments also indicated that the policies tested would be less effective at reducing fuel consumption in a fuel environment exhibit- ing a drastic increase in fuel price or a restriction on fuel availability. This was due to the large mode choice shift to transit, which resulted from the change in the fuel environment. However, a definitive conclu- sion in this area could not be obtained due to limitations in the model- ing system. To Lori ii ACKNOWLEDGEMENTS The author is deeply indebted to his major academic advisor, Dr. William.C. Taylor, Professor and Chairman of Civil Engineering, for his continuous guidance and assistance during the course of this research. The author also wishes to thank Dr. James D. Brogan, Dr. Frank J. Hatfield and Dr. Tapan Datta for their invaluable guidance in the preparation , of this document. Special appreciation is extended to Ms. Vicki Brannan for her help in typing this dissertation. Sincere thanks is extended by the author to all of the members of his family for their continued encouragement and support throughout the course of this effort. iii TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . . . . . . . . LIST OF FIGIJRES o o o o o c o o o o o o o o c o o 0 CHAPTER I - IntrOduc tion . O O O O O O O O O O O O O 1.1 Research Hypotheses . . . . . . . . . . . . 1.2 Goals and Objectives . . . . . . . . . . . . 1.3 Benefits of this Research . . . . . . . . . 1.4 Outline of Report . . . . . . . . . . . . CHAPTER II - Previous Research on variable work Hours and Bus Transit Strategies Related to Fuel Consumption . . . . . . 2.1 Review of Urban Transportation Energy Conservation Strategies . . . . . . . . . . 2.2 Local Bus Supply and Urban Travel Demand . . 2.3 variable work Hour Programs . . . . . . . . 2.3.1 2.3.2 2.3.3 2.3.4 Impacts on work Starting Times . . . . Impacts on Peak Period Transit Demand . Impacts on Peak Period AutomObile Demand Summary of Impacts on the Temporal Distribution of Transit and Automobile Traffic 2.4 Previous Studies on the Simulation of Staggered Wbrk Hour Programs . . . . . . . . 2.5 Justification for the Research . . . . . . . CHAPTER III - The Modeling System . . . . . . . . . 3.1 The Model Requirements . . . . . . . . . . . 3.2 The M003 Model . . . . . . . . . . . . . . . 3.2.1 3.2.2 MOD3 Program Availability and Computer Requirements . . . . . . . . . The Land Use Model . . . . . . . . . . iv wle-‘H 19 25 25 31 31 39 39 42 44 44 46 49 49 3.2.3 Mode Split . . . . . . . . . . . . . . . 3.2.4 Network Assignment . . . . . . . . . . . 3.2.5 Non-Wbrk Trips . . . . . . . . . . . . . 3.2.6 Automobile Fuel Consumption . . . . . . . 3.2.7 Transit Fuel Consumption . . . . . . . . 3.3 Limitations of MOD3 for this Research . . . . 3.4 Modifications of M003 for this Research . . . CHAPTER IV - The Simulation Process . . . . . . . . . 4.1 Introduction . . . . . . . . . . . . . . . . . 4.2 The City Structure . . . . . . . . . . . . . . 4.3 The Base Case . . . . . . . . . . . . . . . . 4.3.1 Base Case WOrk Trip Simulation . . . . . 4.3.2 Results of the Base Case Simulation . . . 4.4 Description of Alternative Policies . . . . . 4.4.1 variable Wbrk Hour Policies (Groups A, B and C) . - . . . . . . . . . 4.4.2 Alternative Transit Policies (Groups D and E) . . . . . . . . . . . . 4.4.3 Combining variable work Hours with Transit Policies (Groups F, G and H) . . 4.5 TWO Alternative Fuel Environments (Policy Groups I and J) . . . . . . . . . . . CHAPTER V - Policy Effects on Fuel Consumption . . . . 5.1 Introduction . . . . . . . . . . . . . . . . . 5.2 Staggered work Hour Programs . . . . . . . . . 5.2.1 Groups A and B: Impacts of the Number of Travelers, and Location of Zones Involved in Staggered WOrk Hour Programs 5.2.2 Group C: Staggering work Hours Along Transit Corridors . . . . . . . . . 5.3 Groups D and E: The Transit Alternatives . . 5.4 Groups F, G and H: Combined Staggered Wbrk Hour and Transit Policies . . . . . . . . 5.4.1 Group F: Staggered work Hours with Transit Redistribution . . . . . . . 5.4.2 Group G: Corridor Staggered work Hours and Bus Redistribution Policies . . . . . 80 80 82 87 93 95 97 97 100 102 103 109 109 110 110 123 123 129 129 134 5.4.3 Group H: Corridor Staggered work Hours and Bus Addition Policies . . . . . . 5.5 Policy Effectiveness in the Alternative Fuel Environments . . . . . . . . . . . . . . . CHAPTER IV - Summary and Conclusions . . . . . . . . . . 6.1 The Policy Alternatives . . . . . . . . . . . . 6.2 Research Limitations . . . . . . . . . . . . . . APPENDIX A APPENDIX B APPENDIX C - APPENDIX D Documentation of MOD3 Program Changes for this Research . . . . . . . . . Data Set for the Generation of the Peak Half Hour Travel for the Base Case . . Detailed Description of Policy Alternatives Congestion Indices for Selected Policy Runs LIST OF REFERENCES . . . . . . . . . . . . . . . . . . . vi 134 138 146 146 149 151 168 182 189 201 Table 1. 10. 11. 12. 13. 14. 15. 16. LIST OF TABLES Techniques Designed to Increase the Effective Processing Capacity of Fixed Capital Transportation Preliminary Estimates of Gasoline Savings from TSM Actions . . . . . . . . . . . . . . . . . Transit Service Headway Elasticities . . . . . . . Source of Riders Attracted to New or Revised Bus Routes . . . . . . . . . . . . . . . . Calibration Coefficients for the Mode Split Model . Coefficients for Fuel Consumption Estimating Relationships for Automobiles of various weights . Input Data for the Generation of the Base Case Activity Pattern . . . . . . . . . . . . Base Case Travel Demand and Energy Consumption . . Summary of Policy Alternatives . . . . . . . . . . Description of Alternative Fuel Environments and Policies Selected for Testing . . . . . . . . . Summary of Bus Policies Tested in the Alternative Fuel Environments . . . . . . . . . Impact of variable work Hour Programs on Wbrk Trip Characteristics . . . . . . . . . . . Effectiveness of Transit Policies . . . . . . . . . Summary of Staggered work Hour and Transit Redistribution Policies (Group F) . . . . . . . . . Effectiveness of Staggered werk Hours Combined with Transit Redistribution (Group F) . . . . . . . Impact of variable WOrk Hour and Transit Redistributions on work Trip Characteristics (Group vii F) Page 15 22 24 57 66 88 96 98 107 108 120 127 130 132 133 Table 17. 18. 19. 20. 21. 22. 23. 24. A1. A2. A3. A4. A5. C1. C2. C3. Summary of Corridor Staggered work Hour and Transit Redistribution Policies (Group G) . . Effectiveness of Corridor Staggered work Hour and Transit Redistribution Policies (Group G) . . Impact of Corridor variable work Hour and Transit Redistributions on work Trip Characteristics (Group G) . . . . . . . . . Summary of Corridor Staggered work Hour and Transit Addition Policy (Group H) . . . . . . Effectiveness of Corridor Staggered work Hour and Transit Addition Policy (Group H) . . . . . . Impact of Corridor variable WOrk Hour and Transit Addition on work Trip Characteristics (Group H) . . . . . . . . . . . . Summary of Policy Impacts in the Alternative Fuel Environments . . . . . . . . Summary of Policy Impacts on Trip Characteristics in the Alternative Fuel Environments . . . . . . Revised Data Input Sequence . . . . . . . . . . . Description of the Additional Input Variables for the Adapted Version of MOD3 . . . . . . . . . Program MAIN Changes for Simulating Staggered work Hours . . . . . . . . . . . . . . Subroutine ZONEWT . . . . . . . . . . . . . . . . Additional Data Requirements for Testing variable work Hours . . . . . . . . . . . variable work Hour Simulation Run Descriptions . Description of Transit-Only Policies . . . . Description of Combined variable work Hour and Transit Policy Alternatives . . . . . . viii 136 137 139 140 141 143 144 152 155 157 158 167 182 184 186 10. 11. 12. 13. 14. 15. LIST OF FIGURES Interrelationship of TSM Action Groups . Matrix Showing Mutually Supportive TSM Techniques Arrayed to Suggest Program Packages Employee Starting and Quitting Times, Downtown Lower Manhattan . . . Arrival Time and Number of Persons Entering the Port Authority Building Lobby Starting Times of All Government Employees in Queen's Park Prior to October 29, 1973 Compared to the Starting Times of Employees on New work Schedules . . . . Changes in work Arrivals and Departures for Bus Passengers . . . . Staggered work Hours -- Passenger Counts at Three Major Downtown Stations Variations in Subway Ridership on the Yonge St. Subway Line in the AM Automobile vo1umes at Screenline B in Ottawa Automobile volumes at Six Ottawa CBD Parking Facilities . . . . . . Basic Requirements of Modeling System Operation of MOD3 . . Relationship Between the Base and Incremental Stage of MOD3 . . . . Causal Structure of Lowry-Type Land Use Model Details of MOD3 Transportation/Land Use Feedback . . . . . . . ix CBD 13 26 28 29 30 32 33 35 36 45 47 50 52 55 Figure 16. ‘17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. MOD3 Average Passenger wait Time as a Function of Transit vehicle Headway . . . . . . . . . . . . . Average Fuel Consumption per Unit Distance versus Average Trip Time per Unit Distance . . . . . . . . Automobile Fuel Economy As Related to vehicle Speed in Urban Traffic . . . . . . . . . . . . . . . Auto Fuel Economy versus Speed Relationships used in NETSIM for Several Acceleration Rates . . . Rate of Cold Start to Fully-warmed-Up Fuel Economy as a Function of Trip Length (Ambient Temperature = 10°C) . . . . . . . . . . . . Diesel Fuel Economy for Transit Buses . . . . . . . Overall Simulation Procedure . . . . . . . . . . . . Zonal Structure of the Simulated Urban Area. . . . . Basic Employment Distribution of the Simulated Urban Area . . . . . . . . . . . . . . . . Highway Network Configuration for the Simulated Urban Area . . . . . . . . . . . . . . . . Transit Network for the Simulated Urban Area . . . . Base Run Land Area (acres) Available for Development . . . . . . . . . . . . . . . . . . Base Run Maximum Residential Density (persons/acre) Population per Zone for the Simulated Urban Area . . Total Employment per Zone for the Simulated urban Area . . . . . . . . . . . . . . . . Temporal Distribution of work Trips Entering the Network for the Base Case . . . . . . . . . . . Zones with a S or 10 Percent Temporal Travel Shift to Time Period 2, and Associated Transit Routes . . . . . . . . . . . . . . . . . . . Zones with a 5 or 10 Percent Temporal Travel Shift to Time Period 4, and Associated Transit Routes . . . . . . . . . . . . . . . . . . . 64 67 69 71 72 81 83 84 85 86 89 90 91 92 94 104 105 Figure 34. 35. 36. 37. 38. 39. 40. 41. 42. A1. A2. A3. D1. D2. D3. D4. DS. D6. D7. Impact of variable work Hour Policies on Total Energy Consumption . . . . . . . . . . . . . Impact of variable work Hour Policies on Total Energy Consumption by Location of the Participating Zones . . . . . . . . . . . . . . . Impact of variable work Hour Policies on Highway Congestion . . . . . . . . . . . . . . . .‘ Relationship between the Highway Congestion Index and Total Energy Consumption for the variable work Hour Policies . . . . . . . . . . . . . . . . Relationship between the Change in the Highway Congestion Index and Bus Ridership Resulting from the variable work Hour Policies . . . . . . . Impact of variable work Hour Policies on Automobile work Trip Travel Time . . . . . . . . . Relationship between Highway Congestion and Automobile work Trip Travel Time . . . . . . . . . Relationship between Automobile Work Trip Travel ,Time and Total Energy Consumption . . . . . . . . Peak Half Hour Congestion Pattern for the Base Case . . . . . . . . . . . . . . . . Algorithm to Factor the work Trip Matrix for Simulating variable work Hour Programs . . . . . . Subroutine ZONEWT . . . . . . . . . . . . . . . . Three Alternative Temporal Distributions of work Trip Travel for the PM Peak Period . . . . . Peak Half Hour Congestion Indices for Run D1 . . . Peak Half Hour Congestion Indices for Runs D2, 03 and D5 . . . . . . . . . . . . . . Peak Half Hour Congestion Indices for Run D4 . . . Peak Half Hour Congestion Indices for Run E1 . . . Peak Half Hour Congestion Indices for Run F1 . . . Peak Half Hour Congestion Indices for Runs C1, G1 and G2 . . . . . . . . . . . . . . Peak Half Hour Congestion Indices for Run F2 . . . xi 113 115 116 118 121 122 124 126 159 161 163 189 190 191 192 193 194 195 Figuue D8. D9. 010. D11. D12. Peak Half Hour Congestion Peak Half Hour Congestion Peak Half Hour Congestion Peak Half Hour Congestion Peak Half Hour Congestion xii Indices for Run F3 Indices for Run F4 Indices for Run F5 Indices for Run F6 Indices for Run H1 198 199 '200 CHAPTER I Introduction One of the many effects of the oil embargo of 1973-1974 on the United States has been a surge of interest in the development of better methods of evaluating transportation plans with specific recognition of energy availability. Current capabilities for planning under energy constraints are Severely limited by the lack of data. During the embargo, the faci- lities necessary to collect gasoline supply and consumption data at a disaggregate level were not available. Since the United States has never experienced a long-term fuel shortage, the influence of such a shortage on urban travel and living patterns is unknown. The influence of transportation system management (TSM) policies to reduce gasoline consumption in urban travel has never been evaluated in an actual urban environment. The only method available for analyzing the potential impact of energy conservation policies or changing fuel environments is extrapolation of observed shorteterm responses through the use of theoretical models. 1.1 Research Hypotheses The primary hypothesis tested by this research was that a reduction in transportation fuel consumption for the urban work trip could be real- ized through the implementation of a staggered or flexible work hour pro- gram. It was also hypothesized that for the short-term a staggered or flexible work hour system could be coordinated with the scheduling/frequency policies of a bus transit system to further improve the savings in trans- portation fuel consumption. For this research study, transportation fuel consumption is defined as the gasoline fuel consumption of automobile work trips plus the daily total diesel fuel consumption of an urban bus transit system. The phrase "short-term" implies a duration of time short enough such that a signifi- cant change in the home to work spatial relationship of urban travel, or the magnitude of urban work travel, would not be realized as a result of the availability or cost of such travel, or as a result of a change in the population in an urban area. In reality, short-term may be several years. For this research the term "variable work hours" did not include four-day work week programs. The variable work hour programs considered in this study were: 1. Staggered work hours -- a fixed schedule of working hours that normally spreads the employee starting and finishing times over a one to three hour period, with individual groups of employees designated to report at 15 to 30 minute intervals. 2. Flexible work hours -- a program where employees start and finish work at times of their own choosing within the constraint that they work a specified minimum number of hours within a given period, and normally within the additional constraint that they be present during certain "core" hours. 1.2 Goals and Objectives The goal of this study was to evaluate the impact of selected vari- able work hour programs and transit scheduling strategies on urban work trip transportation fuel consumption. The work hour programs and transit strategies were tested individually and in combination to allow for the evaluation of the incremental effect of the coordinated policies. The first objective was to establish an operational modeling system to calculate urban work trip tavel demand and fuel consumption for a hypo- thetical urban area. The second objective was to design and evaluate (using the modeling system) alternative variable work hour and transit scheduling strategies. The evaluation was primarily based on a compari- son to a base case of the transportation fuel consumed. The third objec- tive was to demonstrate the possible effect of the prevailing fuel environ- ment on the success or failure of the various strategies to reduce trans- portation fuel consumption. With respect to the latter objective, a "normal" fuel environment was defined as a stable or slowly fluctuating retail price for transportation fuel at levels experienced during 1979. A normal fuel environment also implied the unrestricted availability of transportation fuel to the general public. 1.3 Benefits of this Research The primary benefits of this research were derived from the evaluation of the energy conservation potential of the specific TSM policies tested. The potential for variable work hours or transit scheduling policies to reduce urban work trip energy consumption has never been fully demonstrated, either in the real world or through the use of a modeling system. This demonstration could be a valuable tool to aid in the development of comprehensive transportation plans designed to reduce urban transportation fuel consumption. 1.4 Outline of Report The remainder of this text is divided into five additional chapters. Chapter II describes previous research on TSM policies with specific reference to the energy conservation potential of variable work hours and transit scheduling policies. Based on the literature review, conclusions are drawn for the justification of this research. Chapter III contains a description of the modeling requirements of the research technique, and a detailed description of the modeling system used in this research effort. Chapter IV describes the process used for the evaluation of the alternatives, and also summarizes each individual policy tested. Also contained in Chapter IV is a description of the hypothesized alterna- tive fuel environments used in the analysis. Chapter V contains a de- tailed analysis of the research results disaggregated by policy type and policy combination. Chapter VI is a summary of the major conclusions of the research. Four appendices are also contained in this report. Appendix A in- cludes a detailed description of the modifications required to the com- puter program used for the research. Appendix B contains a complete data set which describes the base case used for analysis in this research when the data is input to the computer program. Appendix C contains all of the data necessary to describe the policy alternatives tested, and Appendix D contains several charts showing the highway congestion patterns and congestion indices for selected policy alternatives. CHAPTER-II Previous Research on Variable work Hours and Bus Transit Strategies Related to Fuel Consumption 2.1 Review of Urban Transportation Energy Conservation Strategies The most widely accepted theoretical modeling system for urban trans- portation planning is the Urban Transportation Planning System (UTPS) developed for the Urban Mass Transportation Administration. Hartgen (1976) evaluated the Capability of the UTPS models to deal with energy constraints. He concluded that the UTPS process can be used to determine the sensitivity of fuel consumption to certain energy policies (for example, speed reduction, increased vehicle efficiency, carpooling). However, Hartgen concluded the process is generally incapable of analyzing the impacts of such policies as rationing or Sunday driving bans. Witkowski and Taylor (1979) suggested that the long-term.conclusions reached by sensitivity analysis would be suspect because traveler reac- tions to an extended fuel shortage would most likely be different than during the 1973-1974 embargo. Therefore, the empirical data collected during the 1973-1974 fuel shortage is not necessarily valid for long- term planning, and hence the relationships developed from this data should not be extrapolated. They concluded that the current emphasis in plan- ning under energy constraints should be placed on short-range energy contingency planning, and methods should be developed to test the impacts of energy related transportation policies. Of particular interest to transportation planners are the evaluation of strategies to reduce automobile fuel consumption in urban areas where approximately 34 percent of the national total transportation energy is consumed (Office of Technology Assessment (1975), p. 3). These trips account for approximately 98 percent of the fuel consumption for urban passenger travel, while supplying transportation for 92-95 percent of the total vehicular person trips (Office of Technology Assessment (1975), p. 17). A review of the energy conservation potential or urban mass transit by the Office of Technology Assessment (1975) concluded that "pure" tran- sit improvement strategies and economic incentives for transit use (includ- ing no-fare transit) can be very effective in attracting increased rider- ship, but they are ineffective by themselves in substantially reducing national energy consumption. It was reported that the most effective methods of energy conservation involve auto use disincentives coupled with transit use incentives. Lutin (1976) reports that, given the current work trip patterns in New Jersey, greater savings in energy could be achieved by using auto- mobiles more efficiently than by increasing public transit patronage. His analysis indicated that carpooling (increasing average auto occupancy from 1.2 to 2.0 passengers per vehicle for the work trip) would save the state 40 percent of the journey-to-work energy. The study also indicated that for present transit operations to contribute an 8.8 percent reduction in work trip fuel consumption, 10 percent of all automobile commuters would have to shift to public transit. This would nearly double 1976 ridership levels. The attractiveness of carpooling and the general ineffectiveness of transit incentives alone to conserve fuel was also demonstrated in a study by Peskin and Schofer (1977). In their study, several hypothetical urban structures were postulated and the impacts of urban transportation 7 and land use policies on transportation energy consumption were evaluated using a simulation model. The results indicated that increasing the work- trip auto occupancy from 1.2 to 1.8 passengers per vehicle would cause a 21 to 42 percent decrease in total urban travel energy consumption de- pending on the city Structure. In contrast to these potential savings, a "free transit" policy resulted in an estimated energy savings of from 4 to 16 percent, while an "express transit" policy did not decrease energy consumption at all. However, under a "no transit" policy, there was as much as a 32 percent increase in total energy consumption, indicating that transit did decrease energy consumption by reducing automobile traf- fic and congestion. This was based on an estimated 13.3 percent of work trips made by transit for the typical bus system simulated. The energy conservation potential of carpooling may be over-estimated in these studies, however, because of the assumed value of increased auto occupancy. Other researchers have concluded that the auto occupancy ob- tainable through carpool incentives is probably in the range of 1.4 to 1.7 passengers per vehicle (Pratt et. al. 1977). In addition, not all carpools represent an energy savings, since, for example, as many as 30 percent of new carpoolers in Portland, Oregon were found to have formerly ridden the bus (Pratt et. al. 1977). Several other studies (Dupree and Pratt (1973), VOorhees (1974), Remak and Rosenbloom (1976), Gross et. al. (1978)) reviewed the potential of different techniques to reduce urban congestion, and to subsequently reduce gasoline consumption. In each of these studies, staggered work hours was regarded as an effective low-cost action to reduce congestion and gasoline consumption. Dupree and Pratt (1973) detailed the findings of a survey and analy- sis of twenty-one low-cost techniques designed to increase the effective processing capacity of fixed capital transportation facilities. A wide range of possible techniques offered promise in satisfying their study objectives. Techniques were rated with particular attention to their potential processing efficiencies (volume increase or time reductions in moving people via existing transportation facilties). In addition, the evaluation considered various cost parameters, impacts on the disad- vantaged, environmental and transportation safety factors, technical and institutional viability, and the expected response from travelers. Techniques were grouped according to their composite ratings and case study analysis candidates were selected from the highest ranked group. The rankings are shown in Table 1. Dupree and Pratt did not review the effects of these techniques in combination or estimate their energy conservation potential. However, staggered work hours was ranked as one of the most promising techniques for increasing highway effectiveness. VOorhees (1974) reviewed the potential of ten "action groups" to reduce urban gasoline consumption. This study attempted to define the interrelationships between these action groups and determine which groups assist, are independent, overlap each other, or are counterproductive. The action groups and these interrelationships are shown in Figure 1. Voorhees indicated that the most significant interrelationships identified in this matrix are: 1. Actions to improve total vehicular traffic flow tend to shift travel from non-auto modes to the automobile, and hence tend to be counterproductive to actions designed to decrease auto travel. 2. Carpooling actions and transit actions are both designed to l) 2) 3) Source: Table 1. Techniques Designed to Increase the Effective Processing Capacity of Fixed Capital Transportation All Around Most Promising Techniques Exclusive Bus Lanes on Urban Arterials (Existing Facilities) Exclusive Reserved Lanes on Freeways for Mass Transit (Existing Facilities) Exclusive Busways on Specially Constructed Rights-of—Way Work Scheduling Changes Highway Traffic Engineering System Improvements Less Generally Promising Techniques Paved Railroad Rights-of-Way High Capacity Transit Buses Organized Commuter Car and Bus Pools Freeway Metering, Monitoring and Control Systems Free or Heavily Subsidized Transit Line Haul Feeder System -Airport Access Improvements Automation of Bus Scheduling Econmmic Penalties and/or Incentives Urban Goods Movement Improvements Para Transit Service (Jitneys, Taxis and Limousines) Least Useful to Achieve Specific Study Objectives The Rail Bus Demand Actuated Transit Service Bus Traffic Signal Preference Systems Auto Driver Aids and Directions Systems The Minicar Dupree and Pratt (1973), p. 6. 10 .mmsouw cowuoc Ema mo mwsmcofiumaouumucH 20312.3 8.8130: 3.083 .e. “mousom .H «Hoodm .mm .m .Avhmflv mwm£u00> < 3.. .ush...u.ouu.”uoz .a .aouoco memo-ion cu nozuo sumo Du ozuoaooLn—uoucaoo «A >3. sou—.3 2395 c0301 .1. so: I $39.0 050.60» 5 unocuzuueuc nausea some any-ovum no mcuuo>o >3 suds: auscuo 530‘ - $39.0 9:02.: ca .3300 some so neat: coca 0.3 :01.) 2595 c032 c . scan-col — Souocc 050:0...» cu eunuo coco undone sous: unseen. c030: '52.... 5:33-3:ch .- , . - . 9693 < _ . 2:2... «Of-‘02 a. Ilka-‘0' .F Zoe—.253 11063 9.3 < - _ _ a...) an :1... .6 54.3.9.8“ .2: yachts: o. coca-col V. .033 < . . < . 2:: s5 a?! x. 8: 9.9.53.3 a. cease-o! to < _ < < . < 0.2.9.1.- sac-ch 003.553.95.103 0 hop—12.90 cl) 1' < - < - - — O LOU 0.10508- 9. IOBIIO‘ .n _ tor. . .0 U 0 _ U U _ , «59-h. scioZo> .18.! 25.15. 0.5.3.1.! .n I I Illalil . Ill-‘ [ll ill I 10‘ IATII. I . . . < . < < < 0 .8010? no! 48065333. 25593 0. connect! .— ua an n; in mmu mu in ma umn mm M m m m. . n m m . . a o .. Wm M a n A A .595 u i s n c d m 29...“: w u a a u N m a. a .o. a u a a we I ... n I m v m m w. W mm m. m .. m m m . a . . m mm m a n a '- 0 (~00 11 reduce the travel of the single occupant vehicle, and hence these two actions may overlap and reduce some of each other's effectiveness. Energy restriction actions and transportation pricing actions act to reduce travel or impose mode shifts in a similar manner and tend to overlap each other's effectiveness. Auto travel disincentives, such as traffic restriction actions, transportation pricing actions, and energy restriction actions, tend to complement and assist incentive actions, such as transit improvements, walk and bike actions, and carpooling programs. VOorhees presented groups of transportation fuel conservation actions for small, medium and large urban areas, with an estimated range of sav- ings for each action individually and for each group as a whole. Measures to improve total vehicular flOW“were estimated to be most effective at reducing fuel consumption for larger urban areas (1,000,000 persons or more), where, for example, staggered work hours were estimated to reduce work trip energy consumption from 1 to 2 percent. Although no figures were given, staggered work hours is suggested to be less effective for medium or small size cities. Characteristics of the implementation of staggered work hours are: Decreased travel time. Minor lifestyle changes. Minor economic impact. Decreased air pollution, noise and congestion. Requires considerable coordination. The energy conservation estimates presented by Voorhees were based on an estimate of the impact of each action on highway congestion and vehicle miles of travel. For the staggered work hours program, estimates 12 were made based on the assumed change in average cruising speed, the number of acceleration/deceleration cycles per mile, the number of stop and go cycles per mile, and the amount of time spent idling. Most of the fuel economy data were taken from a report by Claffey (1971). Remak and Rosenbloom (1976) reviewed the effectiveness of twenty- two techniques for reducing peak period traffic congestion. Actions to reduce congestion would also reduce energy consumption. Five techniques were estimated to be the most effective in reducing congestion in the central business district of large cities: (1) parking controls; (2) road pricing; (3) priority expressway transit treatment; (4) staggered work hours and (5) auto free zones. None of the techniques was found to offer more than marginal reduc- tions in peak period traffic congestion when applied individually. How- ever, packages of effective combinations of congestion-reduction techniques were presented as shown in Figure 2. The staggered work hour package combined five individual techniques: 1. Staggered work hours. 2. Road pricing. 3. Parking controls. 4. Transit marketing. 5. Extended-area transit. The combination of the staggered work hours policy and the transit policies in this package is very similar to the policies tested by this research. Remak and Rosenbloom indicated that, when coordinated with the needs of an existing transit system, staggered work hours have been able to make significant improvements in transit system operating efficiency and increase ridership. 13 m 0 viH u 0 HM o—tu n ~46. H0) 00 o p. m u m at; “'6 -3 0.: was 0 U000) so on Law-‘0 0" -o n o.c u u o NO H ”-4 mu :0 3 >Q°fl¢3 -.c o u u u m x m on an hh0>OH HO to H yum-r4 u-t v4.0 :1: 0.30 H-H tau-nod 3.: c m c u u u u a c o o .o «a ooq o o u a: o o o o > 'F‘ “HN3CCC'Hw-‘HUUH ”'54) U0 Ow-dOhL-o Sim: cccmcumvuuwamjdsm -u o u o o -« u H vgwuuuwumuuuufih OwUOCHWOCO-HHH'UHQM o c c 5.3.“ a are a m m c m e 3 Our-CC Ewooox'oucccic-Ho u c u a ' a u o u m u o m u o x o o 0.4 o o u s u a o-« o u x u m u onmzuaom .H .ONE umouum >c3oco anemone: mfimm .ucoo .mmcwa .psmaom oov.NNO.H mmmtmvo.w mvv.mm ewwm.mmm vhm.bhH.H 0mm.OMH mbfl.¢motm mvv.m® 5mm.mmm .Ou0.mUHmOB A.>H:Uo ocwaommm .umfln .mp0 cumum Hus amouonum mo mcoaaom Hooccov mmwocm oocucouceoz ocm :oHuooHumcou Houoa .Eono .mCAm .mmsm .noom .Hmm ml: mcoHuod Ema Soww mace>om ocflHOmoU mo mounfiaumm muoceawaoum .m canoe 16 II II II II II II II II II II mouom .oxm .mmeo II II II . II II II II II II II mouom .mxm .oaH I- I- I- I- II I- II I- I- I- .am .mem cocoon“; ooh.m 0mm.amv.h II II th.mmh II II vm©.onm hm¢.®o® Nmm.mmh.m mmaoaodm oom.mm voo.wHN II II Ohm.¢v hmatmm hmh.bH II hmhtmfl 0mm.OHH mGMHHu Income 0mm.H II II II II II II II II II .oum .mocmq cuss .oomw A.>flmmo oceaomcm Hence .3050 .mcwm .mmam .soom .Hhm mID .umeo .mmo uuoum Hue mmuuouum mo mooHHcm Hoscoov hmuocm monocoucwcz one coauosuumcou newscaucouc N mHnme 17 mno.~ www.mo II II II II 0mm.mN II II www.5m mom mmoum omm.om mo>.>mm Nvm.a omm.m Nvm.H Nem.H vmm.m vmm.m mmh.h mam.oom .ouu ~ 038 . .usoL II II II II II II II II II II .umom goons omm omm.mam II II II II II mmm.om mma.~m~ II mocom .mom oust II www.mho II II II mmm.mbo II II II II mcwuonm Iooam II oeo.mmm.v Hma omo.HH vmm.maa meo.am hmo.am mmo.m mgm.vv hem.moo.v munch .xm Immo .oom II II II II II II II II II II maaoa II moo.mam.a II oom.mvm II II ovm.omm II ooa.vmm mom.mmv musom .ouo .mmcum omv Nm~.hma II II II onm.ama mmm.v II hoo.ah panacea .mxm .noo . mom .mncoua x.>hawm useeomam Hmuoe .smao Human .mmsm .aoom .usm mus .umao . mo mumum Hue smoumuom mo mooHHom Hmsccmv amuocm oococouowcz one cofluoswumcoo Aemsceucooc N manna mvm.NNN.NN u emz .m .o .Amemfic .Hm .um mmouo "momDOm 18 A.Hmo so ooo.m~m.~ moo.mgh.mm «co. m. mh.a ~¢.~ om.o m.~ h.H om.mm defies ooo.m Hom.vom.o II omm.o omH.mHH II vhm.vv II II bHH.mmm.o mcwuoxucz II II II II II II II II II II mcflnouwcoz coo. 0% II II II II II II II II II mocmcou Ices: mam oao.hmm hmm.omm II II II hmh.> «mg.m II who.H omm.omm Bounmm .mmmm person ¢\z mmm.ama.m II ohH.m om¢.mo 5mm.mm II hoa.o gem.mm vmm.~mm.m .OMcH .mmoL ooo.m oam.o II II II II on.o II II II .Haou one cam.wm , oov.om~.m II II mmm.mvH Hma.~m mvm.mm mmh.m hom.hm www.mmm.h moeuecosd ooo.m~ www.mH II II II II va.mH II How How panache manuscm ooo.mv v¢H.~mm II II II who.aom www.mo II II vom.mmH m a m A.>Msmo unflaonm Hmuoa .Sosu .mcfim .mmsm .soom .Hhm MID .umaa .mwo oucum ewe mmoucnum mo mcoaamm Hmsccmv amuocm ouccooucwcz oco cowuosuuucoo remncaucoo. N magma 19 increased pedestrian activity and bicycling which contradicts statements made by Voorhees (1974). Although some of the studies have addressed the combined impact of several actions, as a whole these efforts made only gross approximations of the impacts of staggered work hours and transit incentives on gasoline consumption. The previous studies did not define the relationships be- tween the size of the participating work force and the impact on fuel consumption, nor did they indicate the magnitude of the temporal redistri- bution of the work trip required to effect a significant reduction in gasoline consumption. There is also little evidence to suggest how tran- sit schedules could be coordinated with the variable work hour programs to reduce energy consumption. 2.2 Local Bus Supply and Urban Travel Demand It is unfortunate that mechanisms were not available during the oil embargo (1973-1974) to capture the information necessary for an extensive analysis of the relationships between transit demand and supply under energy constraints. Therefore, by default, the relationships used in this research were those determined during the more prevalent non-embargo periods. The elasticities between local bus supply and urban travel de- mand under normal conditions probably represents the minimum obtainable headway elasticities under energy constrained conditions, and thus would produce conservative estimates of potential fuel savings. Headway elasti- city is defined as the percent change in ridership resulting from a one percent change in headway. In general, most of the literature that evaluated the effectiveness of improvements in bus transit system operations to induce ridership and/or to reduce work trip energy consumption have drawn conclusions based on 20 conventional operation strategies. Actions to improve transit operations consist of scheduling/frequency changes and bus routing/coverage changes. Scheduling and frequency improvements include: Increasing the number of scheduled transit vehicles. Reducing headways and passenger wait times. Lengthening the hours of service. Rescheduling to provide convenient departure times, or to match regularly scheduled activities, or to provide better coordination at transfer points. An in-depth review of the state-of-the-art of traveler response to transportation system changes has been performed by Pratt et. al. (1977). Traveler response to service frequency changes can vary markedly, however significant conclusions from the Pratt review included: 1. Limited evidence suggests that the median response to frequency improvements is approximately a one half of 1 percent patronage gain per 1 percent frequency increase. Middle and upper income areas are the most sensitive to headway changes provided the prior service was relatively infrequent, (three buses or so per hour) and the trips served are predomi- nantly short. Lower income groups are generally more sensitive to fare changes than to frequency changes, particularly when headways are already short. Sensitivity analysis of bus service to short, suburban trips, done with a mode choice model calibrated for the northern Chicago suburbs, indicated the middle-to-upper income population to be about 40 percent more sensitive to frequency than to fares in the 10 to 20 minute headway range, and over two times more sensitive 21 to frequency than to fares in the 20 to 40 minute headway range (Pratt and Bevin (1971)). Travel demand modeling suggests that transit wait time, plus transfer time, plus walk time may be two to two and one half times as important in mode choice as an equal time spent in the transit vehicle (Quarmby (1967), Shunk and Bouchard (1970)). In various experiments by the Mass Transportation Commission of the Commonwealth of Massachusetts (MTCCM (1964)), bus riders attracted from other travel modes by increased frequency were distributed as follows: Trips Made Previously in own car 18 to 67% in carpool 11 to 29% by train 0 to 11% by taxi O to 7% by walking 0 to 11% Headway elasticities determined from individual Massachusetts demon- stration projects (MTCCM (1964)) are shown in Table 3 along with the elasticities implied by other reported findings (Holland (1974)). The median headway elasticity among those calculated from the Massachusetts experiments is -0.4, or -0.6 omitting depressed urban areas. Actions to improve transit through bus routing and coverage changes include: 1. Introduction of bus service where there was none perviously, or elimination of service where it exists. Major systemwide realignment so as to significantly alter system coverage. Extension of existing routes to provide service to new develop- ments, or other previously unserved areas. 22 Table 3. Transit Service Headway Elasticities Massachusetts Demonstrations, (MTCCM, 1964)— A/ Boston-Milford suburban route (new headway approximately hourly) Uxbridge-Worchester suburban route (new headway hourly) Adams-Williamstown city route (new headway approximately hourly) Pittsfield city route (raised from 3 to 8 round trips daily) Newburyport-Amesbury (depressed area) city route (new headway 30 min. peak/6O midday)§/ Fall River (depressed area) city service (overall 20 percent service increase) Fitchburg—Leominster city route (new afternoon headway 10 minutes, to match morning) B,C/ Boston downtown distributor, Phase 1 (new headway 5 minutes, to match peak) 9/ Boston downtown distributor, Phase 2 (new headway 4 minute base, 8 minute midday) 9/ Boston rapid transit feeder route (new midday headway 5 minutes, to match peak) 9/ Other Reported Findings (Holland, 1974) Study of Milwaukee transit (1955-1970) Detroit city route (new headway 2 minute base, 3-1/2 minute midday) Q/ Chesapeake, Virginia, suburban service 9/ y revenue . B/ Includes impact of minor route extension. 9/ Approximate elasticity computed for full service day by using an unweighted average of peak and off peak (or morning and afternoon) headway improvements. Ey' Arc elasticity calculated by the Handbook authors on the basis of ridership. SOURCE: Pratt, et. al. (1977), p. 160. Headway Months After Elasticity, Implementation -O.4 10-12 -O.2 7- 9 -0.6 1- 3 -0.7 1- 3 -0.4 6- 8 nil 4- 6 -O.3 6- 8 -O.8 5- 7 -0.6 8-10 -O.1 4- 6 -3.8 -- -O.2 -- -0.9 -- Arc elasticity calculated by the Handbook authors on the basis of The by Pratt 1. 23 Initiation of special purpose bus routes to serve specific, inadequately serviced, existing or potential travel demands. Restructuring of a bus system to reationalize service, to accomr modate new travel patterns and to reduce circuitry and the num- ber of transfers required for bus travel. summary of traveler response to changes in bus routing/coverage et. al. (1977) indicated: That the elasticity of patronage to system coverage has been estimated to be in the range of 0.3 to 0.8 percent per 1 percent increase in bus miles of service. Travel models calibrated for Boston indicated transit work trip elasticities of -0.39 for line-haul bus and subway travel time, and -0.71 with respect to changes in transit access time, in- cluding walk, wait and applicable feeder bus travel times. The ability of new or modified bus routes to attract patronage is strongly a function of how a route relates to the local de- velopment, transportation system, and travel patterns. New bus routes have been found to take 1 to 3 years to reach their full patronage potential. The shorter the walk to transit service, the higher the probabi- lity that transit will be used. A 1968 survey of the Buffalo metropolitan area showed that among workers residing 1/10 of a mile from a bus, 20 percent used transit, while among those 1/8 of a mile from a bus, 10 percent used transit. A major component of riders attracted to new or revised bus routes may be riders diverted away from other routes, as shown in Table 4. 24 Table 4. Source of Riders Attracted to New or Revised Bus Routes Radial Routes Circumferential Circumferential Source of to Suburbs Route @ 3 Miles Route @ 5 Miles New Riders St. Louis (97) Boston (176#) Boston (176#) Other Transit Routes 60% 94%5/ 87%2/ Auto 28:9 4% 13% Walk and Other Means 12% 2% u/ New Trips . D/ less than 1% Y .Q/ fi/ 81% of this diversion was from other routes on the same streets. u] 44% other bus routes and 43% rail rapid transit. 9/ 16% single auto driver and 12% carpool. .2/ Not reported. Source: Pratt et. al. (1977), p. 182. 25 2.3 variable work Hour Programs The objective of staggered or flexible work hour programs is to shift work trip travel away from the peak transportation system demand periods. The desired results are a reduction in peak highway and tran- sit system loading, improved transportation levels of service, and re- ductions in energy consumption and vehicle emissions. I Traveler response to staggered work hour programs has been measured at both terminal facilities, such as parking lots or transit stations, and on through facilities, such as highway or transit links. In both cases, the shift in peak period travel demand has been shown to be sig- nificant provided the employee participation rate is high. 2.3.1 Impacts on work Starting Times In April, 1970 the Port Authority of New York and New Jersey in cooperation with the Downtown Lower Manhattan Association (D-LMA), ini- tiated a staggered work hours program to determine the impact of spread- ing out travel demands of workers on public transportation. Prior to the initiation of the program a survey of employee starting times was conducted by the D-LMA. The results shown in Figure 3 indicate that for the 113 firms surveyed (representing 136,000 employees out of a work force of 480,000) 66 percent of the employees were scheduled to start work at 9:00, and 64 percent left work at 17:00 (O'Mally and Selinger (1973)). Initially, 46,000 employees shifted their starting times from 9:00 to 8:30 and left at 16:30 instead of 17:00. Another 4,000 began later, at 9:30 and left at 17:30. The program stressed a shift of at least 30 minutes so as to require a definite change in commuting habits. A dramatic flattening of the peak period arrival pattern occured as exem- plified by the diagram of the temporal distribution of persons entering 26 .NH can NH unnaaxm .xvemac smaflmz.o .mousom .cmuuoncmz Hosea czouozoa .moswa mcwuuwsa one mcfluwoum oomoamsm .m ousmam «7v Si 233% g. -. . III...IIi 0N -fIlllov on oo. .ooxo.a.ew .o 4:332: co_.o_o03< cozoscoi 630.. 550.5500 mmZ: 02.550 mm>0$<0$<mmmm “mousom .Qmo m3wuuo :w m mcflasomnom um mmESHo> magnOEousm .m wusmwm 3.5.. 351' eel. so I 3 3.3.23. 2. a: .23... Ill 0.50.25 go... .8‘ .86 80 On a 8.0 8.0 000 On.» 038? 80 8.. 8. as 85 O p p p h b L p n b n n b p b p h p p b p p L O a u q q . J d. «l q q a a 18 8. Tlllluulllltag: 5...... Tdflfluqnflfln! Tl. u f 1114 1.00 m .1- .IL . HIII IJ _ 0' .II WIIL v T o - 1"J IL . 1 1 18 al-- 1"H"j 8. 8w» . e 1 . f. I II; 1111 M .m a. u.-- ,n . .8 . a. 1. z r .... I III. ”was mun _ :M_ n: 358— 5;. _ . 3 a: 38.—usab- _ .58 33.5.. .uu >8 5.. 832...; 0.3.8) .. 2» Err .. 2» I... s... .52, 36 .HN .m .Amhmav cmwfloz can onw>mwmm ”mousom .mmflufiafionm mcwxumm emu m3nuuo me um mossao> meAOEousm .OH ousmflm nip-.3 pg :8: a 5.358.!» p: 8:335 ‘0 nug‘ 8 I; at»: (Us I 085' .IIIII gusting . 11222 I s 8 C 8.. 8.» 8a a: 3. 3» banana 8. 8.. 3o 8.. R: O p w P w k L1! - w h ¢ b J- T r k b w b m b JP b . .Aalsalsalaaue; vedlndauSllliwi . lg! Ila-III! .11 6a IIIIJ 8L IIUH II T I TL — v A IIJ _ 5' m % FIIL ., . 5 m 0?. — — I“ m n — v I . , no. a 10. I I T 0 fl _ m r L] . _ . III—IIJIIIm Al 3L v I-w-fiII-I—I—v 3 .- .. in». O... “Gib: O... 8.5 .00. .IE 3.. 5' cm. i!‘ Q... .0: '2' .g lb! :0 0g; :0» In: is 0‘; ddbg 37 A staggered work hour program at the 3-M company headquarters in St. Paul, Minnesota was evaluated as being highly successful in reduc- ing peak period traffic volumes at the company industrial complex (Owens and Van WOrmer (1973)). Expressed as a percent of the average daily traffic (ADT), the AM peak 15 minute traffic volume entering the complex was reduced 34 percent, and the AM peak hour volume dropped 25 percent after the program was introduced. During the PM peak period, the 15 min- ute peak period and peak hour volumes decreased 20 percent and 13 percent respectively, when expressed as a percentage of the ADT. The effect was significantly diluted on the highways in the surrounding area. For example, the AM peak 15 minute period and peak hour traffic volumes were reduced nearly 2 percent and 7 percent respectively when expressed as a percen- tage of the ADT on major highways adjacent to the 3-M complex. The implementation of staggered work hour programs has not always been successful. Notably, a staggered work hours plan was adopted by 145 firms throughout the London, England central business district. Only 2 percent of the district work force participated in the program as a result of opposition from business due to anticipated losses in business efficiency. This, coupled with dispersed geographical locations of the participating firms, resulted in a negligible impact on traffic or tran- sit congestion (Dupree and Pratt (1973)). The experience in Atlanta, Georgia (Dupree and Pratt (1973)) dem- onstrated the potential problems inherent in the planning and implemen— tation of a staggered work hours plan. Even though the Atlanta Chamber of Commerce obtained furing for planning and implementing a staggered work hours plan in 1968, no plan has been implemented to date due to strong opposition from business management and employees. Also, failure on the part of the proper authorities to affect transit schedule changes prompted 38 many loweincome employees to resist changes in their work schedules. Traveler response to flexible work hour programs is more difficult to analyze because the timing of work trips is more at the discretion of the employee. The literature review revealed little concrete informa- tion concerning the real impact of flexible work hours on actual work trip travel behavior. However, one study did reach several relevant conclusions concern- ing the impact of flexible work hours on commuting behavior. The Cali- fornia Department of water Resources (DWR) introduced a flexible work hour program in 1974. Based on the responses to a survey of 576 DWR ems ployees, JOnes et. al. (1977) concluded that flex-time is predominantly used to»inprove the match between personal schedules, travel schedules, and work schedules, rather than as an opportunity to change travel mode. However, the responses indicated that flex-time increases the opportun- ity for trial use of buses and carpools, and the direction of experimen— tation was "positive" in terms of the frequency of ride-sharing. It ap- peared that flex-time allowed transit riders and potential riders to match work and bus schedules more closely, absorb a greater degree of unrelia- bility in bus service without suffering the penalty of being late to work, and schedule bus commuting so as to avoid the crush loads and seat- 1ess rides of the peak period. The authors also concluded that while flex-time did not appear to be a powerful incentive for carpooling, its marginal effect did seem to be positive due to an increase in husband- and-wife carpools. Flex-time seems to have improved the quality of the commute trip experienced by many DWR employees regardless of mode. Sixty percent of the respondents reported a favorable impact and 43 percent reported "a major positive impact" in their ability to avoid peak period congestion. 39 2.3.4 Summary of Impacts on the Temporal Distribution of Transit and Automobile Traffic Change in work schedules result in corresponding changes in the peak period travel demand. The changes in the temporal distribution of travel demand appear to vary in magnitude depending on the percent of the work force involved in the variable work hour program, and the proximity of the transportation facilities to the participating employment centers, although neither relationship has been developed quantitatively. The flattening of the distribution of peak period arrivals and departures tends to become less as the distance from the participating employment centers increases. This effect appears due to the increase of "non- participants" in the travel demand patterns with increased distance from participating employment centers. It has been shown experimentally that the percent change in the ob- served magnitude of the peak period automobile and transit demand result- ing from a staggered work hour program will not exceed the percent change in work trip arrivals or departures (Safavian and Mclean (1975)). 2.4 Previous Studies on the Simulation of Staggered Wbrk Hour Programs There have been only a few studies which attempt to determine the impact of staggered work hour programs through simulation of the redis- tribution of work trips during the peak period. None of these studies attempted to translate the results into reduction in fuel consumption. Betz and Supersad (1965) studied the impact of staggered work hours on highway congestion in the central business district (CBD) of a hypo- thetical city. Only the PM (work-to-home) peak period was considered as the authors suggest that the evening movement is the more critical of the two daily traffic peaks. Their method of analysis consisted of an all-or-nothing assignment 40 in five-minute time intervals, the identification of problem intersec- tions (volume/capacity (v/c) ratio greater than 1) and the determination of a possible staggered work hour schedule to eliminate problem inter; sections. First, the intersection with the highest v/c ratio was select- ed, and the largest employment center contributing traffic to this ap- proach was identified. Then the finishing time of this employment center was adjusted within a feasible range to lower the intersection v/c ratios along the paths of the origin-destination pairs for the given employment center. This process was continued for each employment center until all intersection approaches had a v/c ratio less than 1. The base case assumed all employment centers finished work at the same time. The analysis attempted to determine the total amount of stag- ger necessary (in five-minute periods) to reduce congestion at intersec- tion approaches. The traffic assignment in the research by Betz and Supersad did not include capacity restraints. The authors define the "net capacity" of an approach to be the residual from practical capacity after accounting for the non-work peak hour trips, and those trips from small employment locations where quitting times could not be staggered. The "net capacity" was used to calculate the v/c ratio. The results indicated that variations in land use and highway net- work patterns directly affect the efficiency of a staggering system. The dispersed land use pattern showed significantly more intersections with a v/c ratio greater than one when compared to the concentrated land use pattern tested. However, the dispersed pattern showed a lower average v/c ratio. Each pattern required a stagger of seven periods (35 minutes) to relieve congestion for the no-freeway case. The addition of freeways increased the required stagger to 14 and 13 periods for the dispersed O 41 and concentrated patterns respectively. The introduction of freeways tended to increase the problem at those intersections which were already critical, and increased the number and intensity of critical intersec- tions for both land use patterns. It appears that a limiting factor in the efficiency of staggering to relieve urban highway congestion is the operation of expressway ramps and intersections. A study by Santerre (1966) examined the feasibility of implement- ing a staggered work hour program in Houston, Texas and evaluated the potential reduction in freeway traffic as a result of various levels of participation. Santerre surveyed various employment centers to determine the number of employee vehicles using freeway exit ramps near their des- tination. The employee arrival time was also sampled. Using only two employment centers, it was estimated that by removing all of their arri- vals from the peak hour (7:00 AM to 8:00 AM), the average volume measured during ten-minute intervals on the freeway near these locations would be reduced 9.5 percent for the inbound lanes. Santerre concluded that the magnitude of change required to achieve an optimum staggered work hours plan to alleviate morning peak period traffic congestion was not excessive. Tannir (1977) attempted to simulate the effects of a staggered work hour/four day work week program on the operational efficiency of a high— way network serving a high-density employment area in a mediumrsized city. In this study only a very small portion (0.2 percent) of the total vehicle trips using the highway network are included in the simulated staggered work hour program. These trips all had the same work desti- nation. The results indicated the influence of the staggered work hour program was confined to within a two-mile radius of the destination. Given the small number of trips involved in the staggered work hour 42 simulation and several shortcomings in the technique used, the conclu- sion by Tannir that the staggered work hour program realized only margi- nal benefits is less than definitive. 2.5 Justification for the Research The intent of variable work hour programs, from the urban transpor- tation point of view, is to reduce the vehicular traffic congestion and the transit passenger peaking that occurs during the average work day. The objective of lowering and spreading the peak period demand over a longer time period is to increase the efficiency of existing transporta- tion facilities. It is apparent from the literature review that variable work hour programs can be a successful alternative for reducing the congestion of peak period travel. However, the degree to which the reduction in conges— tion can be translated into a savings of transportation fuel is unknown. There are also some contradictory statements made about the potential capabilities of combining variable work hours with improvements in tran- sit service to further enhance the overall level of effectiveness of the TSM strategies. The level of effectiveness of variable work hours appears to be depen- dent on several factors, including: 1. The level of participation in the work force; 2. the relative location of the employment centers participating; 3. the degree of coordination of transit scheduling with the work hours programs: and 4. the highway network configuration. The relationship between these factors and the potential for energy conservation is unknown. If variable work hour programs are to be given 43 serious consideration as a potential means of reducing fuel consumption for the urban work trip, these relationships must be investigated. This research concentrated heavily on factors one through three listed above. The investigation of factor four was not a part of this research effort. CHAPTER III The Modeling System 3.1 The Model Requirements Since the objectives of this research did not include the develop- ment of a modeling system, it was necessary to identify and utilize a modeling system which was capable of accurately simulating the travel impact of the transportation system.management (TSM) policies tested and of estimating the energy requirements of the resultant travel pat- terns. The minimum requirements of the modeling system.are shown in Figure 11. The broken flow lines in Figure 11 represent the feedback mechanism necessary to evaluate the impacts of traffic congestion on mode choice, network assignment, and energy consumption. The capability to evaluate the impacts of congestion or reductions in congestion is the heart of the modeling system. It was assumed that the overall work travel demand patterns were fixed and were unaffected by fluctuations in the cost or time required for travel. The research required a modeling system that included a mode choice model that incorporated the elements of travel time or cost, such as in-vehicle travel time, walk time, and, for transit passengers, waiting time. It was necessary, especially for automobile travel, that in-vehicle travel time be related to highway congestion. In this way the impact on mode choice of increases or decreases in congestion could be shown. It was also necessary that the model be capable of relating changes in mode choice to changes in the frequency of transit service. In this way the effects of frequency changes could be tested along with the variable work hour programs. It was anticipated that the simulation 44 45 Transportation Land Use System System I Travel Patterns I ITrip GenerationI Trip Distribution _______ .._...{Modei Choice I r* l I . t I I | Auto Transit I Travel Travel I I I 1 Traffic (Congestion H: Assi nt Gasoline ‘ Diesel Fuel Consumption Consumption I_Tota1 Energy Consumption _] Figure 11. Basic Requirements of Modeling System. 46 of variable work hour programs and coordinated variations in bus schedul- ing would result in lower fuel consumption. This research also required a model in which the energy consumption for automobile and transit travel were computed separately. This permitted the evaluation of policies affecting both modes, and allowed for the simu- lation of coordinated transit policies and variable work hour programs. The last major requirement of the modeling system was the capabil- ity to simulate the variable work hour programs. That is, the model had to be capable of simulating work travel over several distinct time elements so that the impact of a change in the prOportion of travelers during each time element could be tested. The MOD3 modeling techniques used by Peskin and Schofer (1977) satis- fied most of the requirements of this research, and fit within the limi- tations of the available computer system. The program utilizes an ag- gregation of generally accepted techniques for transportation and land use planning, and was deemed acceptable for use in this research. 3.2 The MOD3 Model MOD3 is based on the modeling structure developed by Edwards (1975) and later extended and modified by Bowman et. al. (1975), and by Peskin and Schofer (1977). Only those portions of MOD3 that are considered relevant to this research effort are discussed. More detail can be ob- tained by referring to earlier research reports Cited above. MOD3 is a large-scale computer model which simulates the spatial development of an urban area, forecasts the passenger travel that takes place during a single day, and computes the energy consumption resulting from that travel. A flow diagram for MOD3 is shown in Figure 12. In effect, the model combines the elements of land use distributions, mode Cihoice, and network assignment with an energy consumption module for 47 Highway Free-Flow Transit Free-Flow Travel Times Travel Times L Initial Relative Interzonal Accessibility Land Use Parameters 1 I I I l Lowry-Type Land Use Model ork Trips Mode Split Based on Previously Defined Highway & Transit LOS Auto Work 7 Trips Capacity-Restrained Transit Free-Flow Equilibrium Assign. Minimum Time Paths Redefined Interzonal Accessibility r——---------—-----v—-——----— Equilibrium Attained? Compute Total Transportation Energy Consumption Figure 12. Operation of MOD3. Source: Peskin and Schofer (1977), p. 16. 48 work trips. The model can also be used to predict energy consumption for non-work trips, but these trips are not part of this research study. The land use portion of this model interacts with the travel portion in three distinct stages to sequentially develop incremental "layers" of land use. The first stage, the base run mode, is the description of the base or existing land use patterns. This description is accom- plished by exogenously defining the basic employment statistics at the zonal level and the highway and.transit network configuration. MOD3 uses a land use model to establish the population and service employment distribution pattern, and then proceeds to develop the initial stage travel patterns and energy consumption. In this stage, changing the values of the input parameters representing travel time or cost, or al- tering the highway or transit network configuration will impact the re- sulting population, service employment and travel distribution for a fixed zonal level of basic employment and a fixed set of land use parameters. The second stage, the incremental run mode, involves the generation of an incremental layer of growth representing the distribution of ad- ditional population and employment for some future date. At this stage, a change in the values of the transportation variables impacts the dis- tribution of population, employment, and travel of only the additional growth layer. The distribution of those elements generated in the base stage remain fixed. However, the mode choice and network assignment is affected by changes in the values of the transport parameters at the incremental level. The final stage represents the future state of the urban area, and is defined by the combination of the base and incremental layers. As originally developed, the policy testing capabilities of MOD3 49 are at the incremental level. .That is, changes in policy are made at the additional growth level to test the impacts on the future develop- ment of the urban area. The relationship between the base and incremen- tal stage of MOD3 is shown in Figure 13. Each stage is defined in a separate computer run, with the combination of the base and incremental stage made during the incremental run to produce the future state. 3.2.1 MOD3 Program Availability and Computer Requirements The Fortran version of MOD3 used in this research and sample data sets were obtained on magnetic tape from the Department of Civil Engineer- ing at Northwestern University in Evanston, Illinois. A printed listing of the program and complete documentation is available in the Doctoral Dissertation by Peskin (1977). MOD3 is a rather large computer program having approximately 3,500 lines of Fortran coding. A typical data set for an incremental run is approximately 1,130 lines in length. Peskin (1977) indicates that the execution time of MOD3 for a typical incremental run was under 65 seconds on an IBM 370/195 computer, with central processor storage requirements under 61,000 decimal words. For this research, MOD3 required approxi- mately 170,000 octal words of centeral processor storage on a CDC Cyber 750 computer, with an execution time under 45 seconds. As a result of the large size of the program and data sets, data manipulation and program execution were performed using a remote terminal. A compiled version of the program and the data sets were stored on mag- netic disks. Printed output was received through a high speed line printer. 3.2.2 The Land Use Model In MOD3, the allocation of land use, population, and service-type employment is based on a Lowry-type (Lowry (1964)) land use model, which 50 r-------.-.-.-.-..---..-.. --“------- ------- ---- --- ---------------- .fi . i I Guinne- . ,- lime ,___T : E Imuelandthm “bflkaiW' I I :lhee III J L-- ------- -“l.---- ----...--------- -..----.----- ”C..-m----H.- P“--------- .I ------------------------------ ------------..--- -01 I i 3 Transfer: I : -'G-IunlimuICbetInsetflnuuome . ; -LendlunanbLe finTZone ; . - annnmmcnaumllhphan-nc ; : lbr1nIII I I I ' I E I 5 I Chang- Addlhufle. I ; Pbkhnr : I I E i I 1‘12““qu ' thermal I ' """ rIt Tri ‘ I had at. I "O W I :5 T : I , I 2 team“: Pneum- I____, : : Lend Ume ' Bork Trum- ; I l 5 5 I I Penna I : lbudknmztrinl : . I I ' I ' * * I I !‘ucure rum : ; (alum-prion . ' i I I I Exam-Intel an: J b------------- ----...------- ...... ”-“-------- ...... --- --- .. Figure 13. Relationship Between the Base and Incremental Stage of MOD3. Source: Peskin and Schofer (1977), p. 50. 51 was developed and calibrated in conjunction with the Pittsburgh Regional Economic Study using data assembled in the Pittsburgh Area Transporta- tion Study. With an exogenously specified set of constraints on available land, acceptable residential densities, and minimum sizes of employment, the Lowry-model allocates the spatial distribution of urban activities. Figure 14 shows the basic structure of the Lowry-type model and depicts the iterative nature of the activity allocation process. The process initially apportions the exogenously specified zonal basic ems ployment to residences based on the relative accessibility of each zone to all other zones. Basic employment is defined as employment in those industries whose products or services depend on markets external to the region under study. Typical of industries that might be considered as basic are manufacturing, national financial institutions, and university employment. The population of each zone is then established by factoring in the labor force participation rate. The level of service employment in each zone is a function of the population of the zone and the rela- tive accessibility. The service employment is allocated to residences and the associated population is again calculated. The resultant in- crease in population requires more services, and the iteration process continues until the incremental population and service employment approach zero. The relative accessibility from zone i to zone j for trip type k is used to allocate population and service employment, and is defined by a gravity model of the form: U k f. .k A 13 rk (1) 13 E U G}? f. .k j J 13 52 IEDBasjmicen I —_.‘ ___J? $I Service] 0 5 lo t . Households N Employmen— (Basic) E ‘ + ———. C Service A Households A Employment (Service) 1? 5 7 7 A II—P Households _._._. (I: ‘ Service‘I (Service) T - - Employment» I Households ‘ —I E II, (Service) S —F L MINIMUM SIZE massaoms] Figure 14. Causal Structure of Lowry-Type Land Use Model. Source: Goldner (1971), p. 101. 53 where A3. = the relative accessibility from zone 1 to zone j for trip 13 type k, Uj = a balance factor in the iterative structure of the Batty (1972) version of the Lowry model, used to prevent over-allocation of workers to residences in zone j, Gj x the exogenously speCified utility or location attraction of zone j for receiving population (for work trips) or service workers (for service trips), k f.. = the interzonal friction factor, a relative measure of the impedance to travel from zone i to zone j for trip type k, a function of the travel time and the dollar cost of travel (including gasoline costs for automobile trips), and r = the trip generation rate per household for trip type k, as- sumed equal to 1.0 for work trips and exogenously specified for service trips. The denominator in equation 1 is the cummulative interzonal access- ability from zone i to all other zones. Therefore, aij can be interpreted as the probability of travel between zones 1 and j from work to home or from home to a service site. The trip interchanges are based on the attractiveness of each zone relative to all other zones in terms of the allocated zonal activities and are computed using a gravity model with all friction factors equal to 1.0, based on the work by Vborhees (1968) for work, social-recreational, and nonhome-based trips. The relationship has the form: P. A. k i 1 Ti" 8 A (2) sz 3 where ij - the trips from zone 1 to zone j for trip category k, Pi - total population in zone i, and {'for work trips: total employment in zone j. A. = 3 for service trips: service employees in zone j. 54 This computation only represents half of the total trips. That is, either the work-to-home portion or the home-to-service site portion of the trip is simulated. The model multiplies the computer trip inter- change by a factor of two to estimate the total daily travel. The interzonal impedance factor, fij' in equation 1 is an integral element in the relationship between transportation and land use. An interative computation of land use is used by MOD3. The highway network congestion depends on the arrangement of land uses which are distributed based on relative accessibility. The essence of the transportation/land use feedback is shown in Figure 15. The generalized cost of travel is defined by a composite of the dollar cost of travel and the dollar value of travel time. The linear function used to compute the generalized cost of travel is defined by: c.. GC.. = t.. + 11 13 13 V (3) where GCij = the generalized cost of travel from zone i to zone j, tij = travel time between zone i and j in minutes, Cij = dollar cost of travel between zones i and j, and V = value of travel time in dollars per minute. The generalized cost of travel is used to compute friction factor values and to define the minimum cost paths on the highway network. 3.2.3 Mode Split The MOD3 program uses a binary logit mode split mode of the form: P =__ei_ (4, ij 1 + eG where P. = the probability of a trip from zone i to zone j by automobile, and 55 4A Free-Flow Highway and Transit Travel Costs ' Exogenous Distribution ‘ - I , of Basic Employment llnitial Mode Split I i e 6 I Lowry-Type Land Use Model Initial Interzonal ; - Friction Factors ' IWOrk Trips] Apply Previously Defined Mode Split [Auto WOrk Trips] Equilibrium Assignment Iterations “0 Final Assignment quired of work Trips Yes ‘L Congested Highway Compute Travel Costs T Non-WOrk Trips [Fevised Mode Split 1 Revised Interzonal * ‘ Friction Factors Figure 15. Details of MOD3 Transportation/Land Use Feedback. Source: Peskin and Schofer (1977), p. 41. 56 e = the base of the nautral logrithm. The exponent G has the form: G = q + r(wA - WT) + s(TA - TT) + t(cA - CT) + u(Own) (5) where q, r, s, t, u = coefficients from model calibration for each trip type. WA, WT = automobile and transit walk times, TA = automobile in-vehicle travel time, including parking, TT = transit in-vehicle travel time, including waiting time, CA = automobile out-of-pocket costs (gasoline and parking), CT = transit out-of-pocket costs (fare and transfer fees), Own = average automobile ownership per household. Talvitie (1972) indicated that the logit formulation yields results comparable to other forms of probabilstic modal choice models. Stopher and Lavender (1972) concluded that the logit formulation was preferrable to other forms of probabilistic mode choice models because the model is less time-consuming to calibrate and less cumbersome to use. The model formulation was ideal for this research because it indivi— dually incorporated several aspects of travel time and cost which relate to mode choice. With this formulation the mode choice was sensitive to both out-of-pocket costs and to the indirect costs of transportation such as congestion and passenger wait time. The values of q, r, s, t and u included in the MOD3 package and used in this research are those from a study by Charles River Associates (1972) for Pittsburgh, Pennsylvania. These values were used in this research primarily because they represent the required coefficients cali- brated to this model structure for a single metropolitan area. It was decided that this consistancy was more important to the modeling process 57 than selecting coefficients from several different studies of smaller urban areas. The values for the calibration coefficients are shown in Table 5 . Table 5. Calibration Coefficients for the Mode Split Model Tri g. r §_ ‘E It: Work Trips -4.8 0.11 -o.o4 -2.2 3.8 Source: Peskin (1977). Included in the transit in-vehicle travel time is a factor represent- ing the average passenger waiting time. This factor is a function of the frequency of transit service and is expressed as: 0.5(6Q/FREQUENCY), Frequency < 4 vehicles/hr. Wait ={ (6) 3.25 + O.25(60/FREQUENCY), Frequency > 4 vehicles/hr. This represents the wait time distribution shown in Figure 16. It.was through this wait time factor that the mode split was made sensitive to transit frequency. 3.2.4 Network Assignment The assignment of automobile trips to the highway network is accom- plished within MOD3 using a capcity-restrained equilibrium assignment algorithm first developed by LeBlanc (1973) and applied by Bowman, et. al. (1975). The computation begins with an all or nothing assignment of automobile trips to the minimum.time (or generalized cost) path between zones. These flows are adjusted by evaluating each link volume and the speed-volume curve to determine the combination of flows which yields the lowest value for the objective function of travel time or cost. The equation used to describe the volume/travel time relationship is given by: 58 20' Wait Time (minutes) g... o) Headway (minutes) Figure 16. MOD3 Average Passenger Wait Time as a Function of Transit Vehicle Headway. 59 T1 = TOi (1 + 0.15 (Vi/Ci)4) ' (7) where Ti = actual link travel time, TOi = link free flow travel time, Vi = assigned link volume, and C1 = link free flow capacity (the volume at which congestion begins). This relationship is the formula used by the Federal Highway Administra- tion for capacity restrained network assignments (COMSIS Corporation (1973)). It is strictly convex increasing, which results in an increase in travel time for each additional unit of volume. The objective function employed by this algorithm to evaluate the optimality of the automobile assignment has the form: . _ 5 . min 2 — E [(1:01) (vi)+(o.2) (T01) (Bi)(Vi)] (8) where Toi and Vi are defined as before, Bi = 0.15/Ci, and C1 is also defined as before. The first term in equation 8 represents the vehicle—minutes of travel on the network at the free flow link travel speeds. The second term represents a portion of the additional vehicle-minutes of network travel that result from the increase in congestion as vehicles are added to the network links. Equation 8 is the integral of equation 7 with respect to link volume, and is evaluated at the assigned link volume, Vi' This algorithm repre- sents a systemPOptimization technique, as opposed to an individual user- optimization, in the sense that the average travel time is minimized 60 (wardrop (1952)). With an individual user-optimization algorithm, travel times on alternative routes between each pair of intersections are equal and are less than the travel time that any individual vehicle would ex- perience by selecting any of the unused routes (wardrop (1952)). ward- rop (1952) indicated that user-optimization is likely in practice since each driver will tend to select a route that will reduce his travel time to a minimum, while systemeoptimization is the most efficient in that it minimizes the total vehicle hours spent. This algorithm ignores the influence of intersections and transit vehicles on traffic congestion, and is only a static assignment technique in the sense that vehicles entering the network during different time elements do not interact. The length of the time segment being simulated must be exogenously speci- fied in terms of the free flow capacity per unit of time for each link of the highway network. For this research the assignment algorithm employed by MOD3 was used to identify the congested traffic links during the peak period. This information was used in turn to identify transit routes where the addition of transit vehicles would result in the greatest reduction in congestion. Transit trips are not specifically assigned to transit routes by MOD3. All transit trips between zone pairs are assumed to travel on the transit minimum time path. Changes in demand on each route can be traced by examining zone pair mode split values, provided there is a unique transit connection between zone pairs. The transit minimum time path algorithm used was developed by le Clerq (1972) and modified by Bowman et. a1. (1975). The algorithm determines interzonal transit travel times by combining parallel route service and weighting the differing route times by their respective 61 frequencies of service. The time spent waiting is based on the combined frequencies of the parallel route services. The resulting travel time includes in-vehicle time, wait time, transfer time, walking time in trans- fering between routes, and walking time between zones. walk time to and from transit at the destination is not included. The algorithm does not include the impacts of automobile congestion nor the time required for passengers to board and alight in determining the minimum transit paths. However, these were not considered to be major weakness for this research. 3.2.5 Non-WOrk Trips. MOD3 simulates the total travel for an urban area for an entire day. In doing so, MOD3 generates and distributes trips for four non- work trip categories including non-home based trips, social trips, and two types of service trips. Wbrk trips are computed considering the effects of congestion on the highway network while non-work trips are not. Hence, only work trips are assigned to the network using the capa- city restrained assignment technique. The simulation of non-work trips was not an integral part of this research. Therefore, the reader is referred to the work by Peskin and Schofer (1977), for details on the formulation of these trip categories. However, the existence of non-work trips in the traffic stream is acknow- ledged through adjustements in the available highway capacity for work trips (see the discussion in Section 3.3). 3.2.6 Automobile Fuel Consumption The computation of automobile fuel consumption in the MOD3 program is processed for each link of the network based on a specified speed/ fuel consumption relationship for both automobiles and transit vehicles. 62 The relationship for automobile gasoline consumption used by Peskin and Schofer (1977) was discarded because the fleet gasoline consumption rates had changed enough to warrant the re-evaluation of the relation- ship. A relationship that not only considered the effect of congestion on fuel consumption, but also reflected the current automobile fleet distribution was more desirable. Using data collected by driving instrumented vehicles in urban traf- fic, Evans et. al. (1976) and Chang et. al. (1977), showed that the fuel consumed per unit distance could be expressed as a linear function of the trip time per unit distance. The relationship is expressed as: ¢=k1+k2t;(x7uau one .Hm .um mcmco .mocmumfla DHCD how mafia mHHB mmmum>¢ mumuw> mocmumwo vac: Hum :ofiumfismcou Hmsm wmmuw>¢ 25:». . uuz< 8v 8m 8w 9: q 1 1 can team: a . 56250 5:92 o m._<.~_m§< 2:2: 0 P [P D P D P bblFL D .’PL "mousom .aa magmas .8. o 2 - fl 8 on R 8— .5.5.. a saw 85:2 5 § § § § mm» o ammo mm 334 oawnsuoa 13m 65 - Suburban cycles with 41.1 mph average speed and 0.4 stops per mile , and - Interstate cycle with 55 mph average speed and zero stops per mile. All vehicles were fully warmed-up before testing. Using the linear relationship for fuel consumption rate in equation 8, the values of k1 and k2 were determined for each automobile. The data from the urban and suburban cycles were combined to approximate the range of speeds used in the study of Chang. The interstate cycle fuel consumption rates were outside the range of linearity. The coefficients k1 and k2 computed for nine automobile weight classes are shown in Table 6. Atherton and Suhrbier (1979) used the results of the Union Oil study to develop a fuel consumption rate estimate for a composite vehicle rep- resenting the average automobile on United States highways in 1976. The resulting values of k1 and k2 are shown in equation 13. ¢ = 0.0425 + 0.01 E gal/mi. (13) where k1 is in gal/mi, k2 is in gal/min, and E is in min/mi. This relationship is plotted in Figure 18 along with the relationship used by Peskin and Schofer (1977) in MOD3. Another technique used to compute the energy consumption of urban traffic is used in the network simulation model (NETSIM) developed by Lieberman et. al. (1977). The NETSIM simulation program is one step higher in level of sophistication for the calculation of fuel consumption rates in that the procedure incorporates the influence of vehicle accel- eration directly into the fuel consumption estimate. The fuel 66 NmHo.O hmao.o vNHo.o omoo.o omoo.o whoo.o mvoo.o mmoo.o vmoo.o AswE\Hmmv me mmmo.o vomo.o mmoo.o omvo.o mmmo.o hmmo.o mamo.o m©N0.0 Mbmo.o Awe\ammv ax oomm ooom oomv 000v comm Doom mmom ommm ooom .mn< .m .Amhmav uoflnusom osm couumnud "mousom omhm ommm omhv ommv omhm ommm mhmm mhmm mNHN Amado xumz mmmHo N H m x + x u e imbue mmmao HmNm m Hmhv 0H HmNV ON Hmhm NH HmNm NH QBQN NH 05mm 5 QNHN m mhmH m unmwms Ilmmmmmwl moaoano> mo awnssz munmao3 moofium> mo moawnosousr How mmwnmcowumHom moguMSwumm :oeumfismcoo Hash How mucoaowwwooo .o magma 67 .owmmmufi coon: ca omomm oaow£o>.ou ooumaom mm >Eocoom Hash oaflnofiousm .ma ousmwm Auden Ham moHflEv ommmm cm or om om on on om 0H 0 q Jl « a u u q d \LN no no . m 0H we cowumsoo nma cowumsaxoummm szBmz Iva ma cowumoUo 0H oaowro> ouwmomsou 1 moo: Hmsfimwuo ma Homonom pom sermon N N N (uotteb 19d setrm) Kmouooa tang 68 consumption/speed-acceleration relationship used in NETSIM is shown in Figure 19. In two separate reports (Evans and Herman (1978), and Evans (1979)) it has been shown that the simulated fuel consumption rates generated by NETSIM can be approximated accurately by a relationship of the form of equation 8. Evans and Herman (1978) calibrated equation 8 to data generated in a simulation study by the Honeywell Traffic Management Center (1976) for a network of 34 intersections. The results were: 0 = 0.047 + 0.0133 E gal/mi (17 < ~' 35 mi/hr.) (14) where E is in min/mi. The values of k1 and k2 used in equation 14 yielded an r2 (square of the correlation coefficient) value of 0.998 and a root mean square error in prediction of 2.21 when compared to the Honeywell data. Evans (1979) also showed that equation 14 closely approximated the NETSIM fuel consumption statistics reported by Christopherson and Olafson (1978). The values of k1 and k2 in equation 14 also compare favorably to the values in equation 13 generated for the 1976 composite vehicle. The relationship in equation 14 is also plotted in Figure 18. It can be seen in Figure 18 that the relationship originally used by Peskin and Schofer (1977) in MOD3 uses slightly higher fuel consumption rates than either equation 13 or equation 14 for speeds less than 12 mph and 20 mph respectively. Equation 13 uses a lower fuel consumption rate than equation 14 for the entire range of speeds. However, it should be noted that equation 13 represents the fuel consumption relationship for a fully warmed-up vehicle. Atherton and Suhrbier (1979) report that cold- start condition, typical of most urban work trips, reduces fuel economy 69 .v .m .Ahhmdv pawwwswmom ocm coaumnwwa ”mouoom .mwumm cofiumuwaooofl Hmuw>mm new szBmz a“ coma mmwnmsoflumamm oommm msmuw> hfiocoom Hush ou=< Aumm\hmv amwam a a a s a a a Ammauv zo_bm3£ofim .mm musmfih .>63 0:0. m~ Ea. oommm touh ovum xcqq 4 on o no a com.— oom.~ ooo.~ ooo.~ com.— arm». >u«ooaau stud m v n M mn~.o I .oo.o..om.o..om.o. I u~u<¢h we“: o.oo u mean» xuoa new o~na~4o>o >uuuenau assumes om.o . ..6u:. coaaaaanau .6 uoauoa onae o.om I use: non use» sauna accouon 86 Figure 26. Transit Network for the Simulated Urban Area. 87 used by Peskin and Schofer (1977), and is representative of urban bus routes in United States cities in terms of route spacing and average link speeds. The focal point of the network was the CBD, and the network was designed such that each zone had access to transit. All routes began and ended at the city periphery. Where possible, the use of multiple routes serving any single zone was avoided to enhance the capability of monitor- ing changes in transit ridership that resulted from changes in route frequency. 4.3 The Base Case The base case, used as the control to compare the effectiveness of alternative TSM policies, was generated using the MOD3 program first in the base run mode, and then with one run in the incremental run mode with the model constrained by a "no-growth" situation (see Chapter III for model description). The difference between the base run mode and incre- mental run mode is that, in the former, travel time was used to compute the interzonal friction factors and define the minimum paths in the high- way network while in the latter, the generalized cost of travel was used. This resulted in a slightly different activity pattern when the base run output was used in the incremental run mode even though a no-growth situa- tion was specified. All policy test runs were made in the incremental run mode (no-growth) to avoid the redistribution of urban activities from the policy alternative. Table 7 and Figures 27 and 28 give the values of some of the input data required to describe the base activity pattern of the study area. A complete listing of the data set for the base case is shown in Appendix B. With this information as input, the resultant population and employment distributions are shown in Figures 29 and 30. 88 Table 7. Input Data for the Generation of the - Base Case Activity Pattern Percentage of persons working at home Value of traVel time for work trips Price of gasoline per gallon Automobile Occupancy rate for work trips Automobile Ownership per household Parking cost - CBD: Wbrk trips Per day Non work trips Ring 1: WOrk trips Non work trips Elsewhere: Number of transit routes Peak period transit frequency of service (buses per hour) Transit bus trips per day on each route Transit fare Transit transfer fare Labor force participation rate (persons per employee) 2.3% $5.00 $1.00 1.3 1.3 $2.50 $1.00 $1.25 $ .50 Free 12 43 $ .35 $0.00 2.5 Base travel time distributions - Supplied by Peskin and Schofer (1977), originally taken from a Texas Department of Highways (1964), study of Amarillo, Texas. Acres of land available for development in each zone (see Figure 27). Maximum Residential Density, in persons per acre per zone (see Figure 28). 89 1505 565 376 376 376 565 565 494 329 329 329 494 565 329 329 173 173 173 329 376 1505 329 329 173 5° 173 329 376 1505 329 329 173 173 173 329 376 494 494 329 329 329 494 565 565 376 376 376 565 H 1 mile N 1505 Figure 27. Base Run Land Area (acres) Available for Development. 9C) 9.4 13.3 13.3 13.3 13.3 13.3 9.4 9.4 13.3 16.6 16.6 16.6 9.4 - 9.4 uzd 6.3 9.4 13.3 18.8r': 16.6 9.4 9.4 6.3 9.4 13.3 16.6 16.6 16.6 9.4 9.4 9.4 9.4 9.4 9.4 9.4 9.4 9.4 9.4 9.4 9.4 9.4 9.4 1 mile Figure 28. Base Run Maximum Residential Density (persons/acre). 91 950 Total - 100,000 900 1066 1196 1072 926 763 2216 2737 3166 3539 2162 650 961 3694 3206 3322 3270 2664 1057 674 1070 2959 3370 84 3349 3060 1136 , 1024 6 994 2707 3165 3394 3165 3521 1113 603 2066 4071 3026 2963 2324 933 696 1106 1142 1076 916 1 mile N 1002 Figure 29. Population per Zone for the Simulated Urban Area. €32 345 16651 - 40.000 316 365 349 347 301 292 992 643 656 654 966 314 334 653 1626 1632 1632 640 363 337 335 655 1636mm 1635 656 346 346 352 637 1625 1640 1626 657 349 309 995 613 670 654 999 306 302 351 352 369 320 H 1 mile N 349 Figure 30. Total Employment per Zone for the Simulated Urban Area. 93 4.3.1 Base Case WOrk Trip Simulation To facilitate the testing of staggered and flexible work hour pro- grams, the total PM peak travel period was segmented into five discrete time elements. The work trip travel for each time element was then simu- lated using the adaptations made to MOD3 for this purpose. The sum of the energy consumed during these five time elements represented the total for the entire peak period. The peak travel period was specified to have a length of 2.5 hours and was divided into the five half hour periods. Half hour time periods were selected for three basic reasons. First, it was felt that half hour periods were adequate to describe the peaking characteristics of urban work travel. Simulating more time periods of a smaller duration would have resulted in only a small increase in descriptive capability at a substantial in- crease in computer costs. Second, the work by O'Malley and Selinger (1973) stressed that a travel time period shift of at least 30 minutes was neces- sary with a variable work hour program to require a definite change in commuting habits. Third, the use of half hour time periods eliminated the potential problem of vehicles from different time periods interacting on the network. Which was a condition the simulation process was incapable of accounting for. The mean work trip travel time for a city of the size simulated is about 8 to 10 minutes (Peskin and Schofer (1977)). The base case temporal distribution of evening work travel is shown in Figure 31. The general shape of the distribution is similar to the distributions found in studies of urban work trips (Port Authority of New York and New Jersey (1977), Santerre (1966)), although the peaking characteristic of the base case is slightly less exaggerated than than found in the literature. It was found that loading the simulated network with more than 50 percent of the total work trips during a half hour Percent of Total Trips Entering the Network 50 45 40 35 30 25 20 15 10 94 1 4 1 L . 1 l 2 3 4 5 Half-Hour Time Period Figure 31. Temporal Distribution of WOrk Trips Entering the Network for the Base Case. 95 period resulted in unrealistically high levels of congestions. Each half hour time period was simulated to determine the levels of network conges- tions, transit ridership and energy consumption for the work trips. The highway congestion index (HCI) reported by MOD3 was used as a measure of average congestion on the entire network. The HCI is the mean of all the congestion indices computed for each link of the network. The congestion index for each link is defined as the ratio of the link free flow travel speed to the link travel speed when adjusted by the link volume of traffic. As the level of congestion increases, so does the HCI} 4.3.2 Results of the Base Case Simulation The results of the base case simulation for each time period, and for the entire peak period are shown in Table 8. The last column in Table 8 contains the weighted mean values of the travel and congestion indices for the peak period. The mean values for the entire peak period were computed by summing the weighted mean values for each time element. For example, the mean highway congestion index reported for each time element was combined into a mean for the entire peak period based on the number of automobile trips occuring in each time element. That is, n wti = * Hume;m .2 WT hci (15) i=1 where HCI = the mean highway congestion index for the entire peak period, mean n = the number of time elements in the peak period, wt. = the number of automobile work trips occuring in time period i, P n WT = the total number of automobile work trips = Z wti, and i=1 hCi = the mean highway congestion index for the time period i. 96 GOeumEsmcou hmumcm om~.m mmH.m bmH.N Hm~.m bma.m oma.m AAMEV ommmm mom.hm mmh.mm oom.wm Hom.mm vom.om mmh.mm Ammuncefiv mafia oov.a oom.H va.H, Noe.H me.H ovm.a Amoaflsv numcwq "mews xuo3 Hemcmua monum>¢ omh.ha 00m.HN vmo.a~ mum.va mmo.Hm mem.am Asmfiv 000mm NmH.m vmm.h Ohm.h mmm.oa Ohm.h vmm.h Amouscwsv mafia omo.m th.m Hmo.m mmm.~ Hmo.~ mum.~ Amoaesv spaced "mega xuoz owed ommuo>¢ mmv.vH moo.MH . Nmo.mH www.ma Nmo.ma moo.MH mmeua xuoz «0 wooonmm 50¢.NH NNH.H mmo.a Hmh.o om0.H NNH.H mmwue cOmHom annexe xuoz nemomwa H6009 mmm.om ~ee.m omo.m ~mo.e~ mmo.m mee.m maaue 606966 xuoz 095a mvm.a moo.H mmo.a 5mm.m mmo.H moo.H woocH sowumomcoo zmznmfim H6009 www.mva iooa x 6.55mi amuwcm mam edema Hmuoa hmm.hmh.v nmn.mao www.mmm Hmm.mmo.m www.mmo www.mav Aooa x miDBmv wmuosm mega xnoz ouom Ammeua xuoz Hmuoe mo :oeuuomoumv oa.o mH.o om.o. mH.o oa.o COeuanfiuumeo panama ooeumm room :66: m e m m H ooaumm meme teammaoz one confion Hm>mna mmmu wmmm .m manna 97 This was done for all of the summary statistics, except for the auto- mobile and transit energy consumption. The total automobile energy con- sumption is simply the summation over all time periods, and the transit energy consumption is a daily total. 4.4 Description of Alternative Policies The policy alternatives tested by this research were divided into three basic categories: 1. Variable work hour policy only, 2. transit policies only, and 3. combined variable work hour and transit policies. The work hour and transit policies were tested individually and in come bination to determine whether or not the combined policy alternatives could further improve the savings in energy consumption of the individual policies. Within each policy category, individual policy alternatives were grouped based on the underlying structure of the policies. These groups are summarized in Table 9. Each group was designed to test specific para- meters relating to policy effectiveness. Detailed descriptions of each group are contained in the sections that follow. 4.4.1 Variable WOrk Hour Policies (Groups A, B and C) The simulation of variable work hour programs was accomplished by changing the temporal distribution of work trips entering the network. Several variations were simulated to test the impact of both the number of participants in the program and the location of the participants. The variations tested represented the potential impact of both staggered and flexible work hour programs for a five-day work week. For all the policies tested, workers were assumed to remain on a five-day work schedule. 96' Table 9. Summary of Policy Alternatives variable WOrk Hour Alternatives Group A Basic Policy Structure: Shift travelers away from the peak half hour. vary magnitude of shift by 10, 30, 50 and 60 percent of peak half hour. Vary zones involved. Group B Apply total temporal distribution of work travel resulting from Group A policies to all zones in the study area. Group C Apply 10 percent shift to zones along east-west and north-south transit corridors for different time periods. Transit . Alternatives Group D Transit Bus Redistri- bution with base Temporal Distribution of work trips. Group E Addition of 6 transit buses to network, with base temporal distribution of work trips. Combined variable Wbrk Hour and Transit .Alternatives Group F Combine Group A with Group D policies Group G Combine Group C with Group D policies. -Group H Combine Group C with Group E policies. 99 The basic structure for testing the staggered work hour programs was to shift work travelers away from the peak half hour to each adja- cent half hour period. This was continued (incrementally) until the adja- cent half hour periods each contained approximately 20 percent of the work travel for the zones under study. Additional travelers were then shifted to time periods on both sides of the peak, until a uniform distri- bution was established (that is, 20 percent of travel during each time period). The shifts were made in increments of 10, 30, 50 and 60 percent of the peak period travelers. The 60 percent participation rate for the peak period travelers equaled a uniform distribution for the zones in- volved in the travel time shift. To test the impact of the location of the participants, as well as the overall number of travelers, the simulation began with only the CBD zones participating, and progressed outward from the CBD adding adjacent rings of zones in successive program runs. In all, seventeen variations of the basic staggered work hour programs were tested while maintaining the base transit frequency (runs Al-A17). Table 8 shows the zones in- volved in the staggered work hour program along with the temporal distri- bution of all travelers from all zones. These alternatives are described in tabular form in Table C1 in Appendix C. For five cases (runs Bl-B5) the overall temporal distribution of work travel that resulted from the staggered work hour simulations for selected zones (runs A4, A5, A8, A9 and A12) were applied to all zones. This was done to test the impact of concentrating the staggered work hour program in selected zones as opposed to generating the same overall tem- poral distribution of travelers over all zones. These runs are also des- cribed in Table C1. 100 A variation of the staggered work hour policy was designed to coor- dinate the staggered work hour shift along selected transit corridors. In run Cl, 10 percent of the travelers originating in the zones along the transit routes which traverse a general eastdwest direction were shift- ed from time period 3 (the peak period) to time period 2. The same per- centage of travelers originating in zones along transit routes traversing a general north-south direction were shifted from time period 3 to time period 4. The purpose of this variation was to attempt to enhance the influence of the transit system on work travelers involved in the variable work hour program. This run is also described in Table C1. The policy structure described for run C1 was also used as a basis for testing com- bined staggered work hour and transit policies. 4.4.2 Alternative-Transit Policies (Groups D and E) The base case transit frequency of service was set at 4.5 vehicles per hour (13.33 minute headways) for the peak period. An analysis of the total travel time on each route revealed that this frequency could be maintained utilizing 3 buses per route, yielding a total fleet require- ment of 36 buses. The total number of trips per day per route, which was used in the computation of the daily transit energy consumption, was set at 43 for the base case. Two groups of transit policies were tested in this research. The first group assumed that the fleet size was fixed and that any increases in frequency requiring the addition of buses to a route would necessitate a decrease in frequency on one or more routes. Analysis of the automobile network assignment and subsequent congestion patterns suggested ways in which the transit fleet could be redistributed to reduce congestion and energy consumption during specific time periods. Several combinations 101 of fleet redistribution were simulated using the base case temporal dis- tribution of work trips. These alternatives are described in Table C2 (Appendix C), where, for example, in run Dl one bus was added to each of routes 2, 3, 8 and 9 during the peak half hour, increasing the bus frequency on these routes to 6 buses per hour for that half hour period. These buses were taken from routes 1, 4, 7 and 10, reducing the bus fre- quency to 3 buses per hour for the period. The purpose of this exercise was to determine if it was possible to reduce energy consumption through a redistribution of the transit fleet. .Each redistribution of the fleet also resulted in a change in the total number of daily transit trips made on each route. Evaluation of the impact of each alternative on energy consumption and congestion was used as feedback to suggest a means of gaining further improvement (see Chapter V). In all, five such policy alternatives were tested (runs Dl-DS). The manual evaluation of the congestion patterns and interpretation of possible improvements was to illustrate the impact of these changes, and was not meant to produce an optimum distribution of the fleet. In terms of a reduction in energy consumption, the transit redistri- bution policies provided only a slight improvement (see Chapter V). The redistribution policies severely penalized those routes where bus frequency was reduced with a loss in ridership, and therefore restricted the total benefit realized by the redistribution policies. As a result, a second type of transit policy was designed. This policy consisted of using 6 additional buses on the transit network. Only one bus addition policy was tested. This alternative is described in Table C2 as run El. 102 4.4.3 Combining variable work Hours with Transit Policies (Groups F, G and H) ‘ To evaluate the combined impact of the transit policies and the var- iable work hour programs, several alternatives were designed that included both a transit redistribution (or addition of vehicles), and a variable, work hour alternative. In all, nine combined alternatives were simulated: six combined transit redistribution with the basic variable work hour policies (Group F), two combined transit redistribution policies with the transit corridor oriented variable work hour policy specifically de- signed to enhance transit ridership during the time periods adjacent to the peak half hour (Group G), and one combined the addition of transit vehicles with the transit corridor variable work hour policy (Group H). The policy combinations in Group F were an attempt to determine the' reduction in energy consumption of the basic variable work hour policies by combining them with the most successful transit redistribution poli- cies. For example, run F1 combines the temporal distribution of work travel from run A2 with the transit bus distribution of run D5. However, feedback from the evaluation of these simulations indicated that changes in the bus redistribution strategies might further improve the energy conservation potential of an alternative. Therefore, slight deviations from the transit policies tested with the base travel distribution were made. These policies are described in detail in Table C3 in Appendix C. The variable work hour policies used in combination in Group F were for both the 10 and 30 percent participation rates for the peak half hour. The location of the zones involved varied from strictly the CBD zones to all zones from the CBD to ring 3 of the urban area. The variation of the zone location for those zones involved in the variable work hour program resulted in a shift of traffic congestion on the network, and 103 precipitated the re-evaluation of the transit distribution strategies tested in Group D (see Chapter V). The policy combinations simulated in Groups G and H had, in effect, a common goal. These policies attempted to enhance the attractiveness of transit for work trips by shifting some of the work trips along the east-west and north-south travel corridors away from the peak half hour, and then improving transit frequency along the routes where the shifts were made. For example, Figure 32 shows the zones that experienced a 5 or 10 percent variable work hour shift to time period 2, and also shows the transit routes with an increased frequency of service during this time period. Figure 33 shows the same information for time period 4. The alternatives in Group G combine the corridor staggered work hour policies with transit redistribution policies. Group H combined the corridor staggered work hour policy with the transit addition policy. The results of all policies tested are described in detail in Chapter V. 4.5 Two Alternative Fuel Environments (Policy Groups I and J) The 1973-1974 Arab oil embargo of the United States caused signifi- cant changes in work trip travel habits in many urban areas as a direct result of limited fuel availability and higher fuel prices. It has been estimated that these changes could include a potential 40 percent increase in daily transit ridership (Office of Technology Assessment (1975)), the majority of which would occur during peak hours due to work travelers changing mode. It is likely that the effect of TSM policies would be significantly different in an environment exhibiting rapidly increasing fuel prices and/or a limitation on fuel availability. Two alternative fuel environments and several policies were formulated to test this theory 104 10 - 10 percent shift 5- 5 percent shift IO 1 1 10 Q)... ______ .-, 5? I I 10 10 IO 10 1 10 10 I ”’r\ ' / 10 10 5 s 10 i 10 / L- ‘ \ 43> {a —— ---.-Jk:“ ’fip- n----4h--J :— VI 1” I 10 V5 10 ’40 ' L--1D-dD---- -.v I i 10 10 10 10 I © 5. ----- _--® Figure 32. 1 mile Zones with a 5 or 10 Percent Temporal Travel Shift to Time Period 2, and Associated Transit Routes. 105 Figure 33 . 10 \\ 10 - 10 percent shift 10 1° 1 \ 1° 1° 5 - 5 percent shift ®.-.qp--uh-1‘l \ ? II I ! l 1c I10 [ 10 I I I . ! s 10 "1o _ .I-J .. T 50 I 5 I'M, " l v v‘ 10 J10 2‘15 r- p - I T l I ' 10 I10 '10 I . I I I I I I J I I i 10 :10 :10 10 10 I I \ \ H \ 10 1 mile N \\ Zones with a 5 or 10 Percent Temporal Travel Shift to Time Period 4, and Associated Transit Routes. 106 using the MOD3 program. The energy environment of the simulated urban area was altered drastically through two measures. The first was to raise the price of gasoline from $1.00 per gallon to $3.00 to represent a possible price reaction to another oil embargo (Group I policies). A second change was simulated by reducing by 50 percent the coefficient of the automobile/ bus travel time differential in the work trip mode split relationship in combination with the increased gasoline price (Group J policies). Hence, the relative importance of travel time in mode choice was reduced by half. This represent a possible user reaction to a severe limitation on fuel availability. These changes were tested initially using the base case temporal distribution of work travelers. Combinations of policies were then tested beginning with the corridor staggered work hour policy (identical to that described for run C1) and including a bus addition policy similar to that described for run H1. The bus addition policy was nearly identical to that described for run H1 in that during time periods 2 and 4 one bus was added to each route that was coordinated with the corridor staggered work hour policy. A slightly different distribution of the additional buses was indicated for time period 3 by the analysis of the individual link congestion indices as described earlier. These policies and the prevailing fuel environments are summarized in Tables 10 and 11. Table 10. Number I1 I2 I3 J1 J2 J3 Fuel Environment, Gas $3.00/ga1. Unlimited Avail- ability Gas $3.00/ga1. Unlimited Avail- ability Gas $3.00/gal. Unlimited Avail- ability Gas $3.00/gal. Limited Availability* Gas $3.00/gal. Limited Availability* Gas $3.00/ga1. Limited Availability* 107 Temporal Distribution of work Travelers Base Case Corridor Staggered work hours (See run C1) Corridor Staggered work hours (See run C1) Base Case Corridor Staggered work hours (See run C1) Corridor Staggered work hours (See run Cl) Description of Alternative Fuel Environments and -Policies Selected for Testing BUS Policy Base Case Base Case Bus Addition (See Table 11) Base Case Base Case Bus Addition (See Table 11) * The impact of limited fuel availability on work trip mode choice was simulated by reducing the coefficient of the automobile/bus differen- tial travel time by 50 percent in the mode split relationship. 108 Table 11. Summary of Bus Policies Tested in the Alternative Fuel Environments Routes Number of Run Receiving Buses Added ~Time Numbers Buses Per Route Period I3 and J3 1, 7 1 2 5, 11 6, 12 1, 2 3 3, 9 6, 12 2, 8 1 4 3, 4, 10' CHAPTER V Policy Effects on Fuel Consumption . 5.1 Introduction Each of the policy alternatives described in Chapter IV was tested using the MOD3 program in the incremental mode with a no-growth situation specified. The primary measure of effectivenss was the relative change in work trip energy consumption with respect to the base case. Other evaluations were made with specific alternatives as the basis for compa- rison, for example, where a combination of policies was evaluated against each individual policy comprising the combination. The total work trip energy calculation contains both the automobile and transit consumption data for an entire day's travel. Transit energy consumption is computed by MOD3 as a daily total and as such the contri- bution of transit energy consumption from each individual time element cannot be specified. However, this is not a major drawback in the analysis. Automobile energy consumption was reported for each individual time element and was a useful tool for the analysis of policy impacts. For the base case, the total daily transit energy consumption was only 3 percent of the combined transit and automobile work trip energy consumption. For non-transit policies, the inclusion of the daily transit energy consumption resulted in a maximum difference of only 0.3 percent when evaluating the policy impacts on energy use. This occurred in con- junction with the staggered work hour policy that resulted in the largest percentage decrease in energy consumption. The difference was much smaller (approximately 0.1 percent) for non-transit policies, resulting in a lesser decrease in total energy consumption. Combining transit and automobile work trip energy consumption facilitated the direct comparison of transit 109 110 and non-transit policies. 5.2 Staggered work Hour Programs The results of the staggered work hour alternatives simulated in this research can be interpreted to represent the potential of such alter- natives to relieve congestion and reduce energy consumption for the city simulated. The variable work hour policies can be interpreted to repre- sent either a staggered or a flexible work hour program, or a combination of both. The importance of the results resides in the magnitude of the shift obtained, not in the mechanism for obtaining the shift. For con- venience, the policies are referred to as staggered work hours. The total energy consumption resulting from the staggered work hour policies was influenced by two primary parameters -- the total percent of the work force involved in the program and the location of the zones participating in the shift with respect to the CBD and the transit network. 5.2.1 Groups A and B: Impacts of the Number of Travelers, and Location of Zones Involved in Staggered Werk Hour Programs The results of the simulation of the staggered work hour programs on automobile work trip and daily transit energy consumption (hereafter referred to as total energy consumption) are shown in Figure 34. The results show that there is a strong relationship between the percentage of work travelers shifting away from the peak half hour period and the percentage decrease in total energy consumption. The smooth curve shown was manually fitted to the data, and represents the approximate relation- ship between work trip travel time shift and the potential energy savings. This relationship asymptotically approaches a 12.2 percent energy savings for work travel at the point where the temporal distribution of work travel is uniform over the length of the peak period. 111 The curve in Figure 34 indicates that the potential for energy sav- ings from staggered work hour programs is much higher, for a city of the size simulated, than is indicated by the estimate made by VOorhees (1974), which was less than one percent. A 10 percent shift of work travelers away from the peak half hour resulted in a 4 percent savings in energy. A 10 percent shift can be considered a realistic goal for such a program based on the program results reported by O'Mally and Selinger (1973). In their study of the impacts of staggered work hours in downtown Manhattan it was reported that approximately 10 percent (46,000 employees) of the total CBD work force participated in a coordinated program from 113 dif- ferent firms. Safavian and Mclean (1975) reported a 50 percent partici- pation rate (35,000 employees) in a program in Ottawa, Ontario. However, all of these were federal government employees. The data from this research also indicated that it is better from an energy standpoint to implement programs that have a dispersed area of influence rather than concentrating the effort in a small area. For example, simulation runs Bl and BS resulted in a larger energy savings than runs A5 and A4 respectively. Runs B1 and BS had the same total work trip shift from the peak periods as runs A5 and A4 but the shift was ap- plied to the entire urban area rather than concentrating it in the CBD. This difference dissipates as the total base area of influence becomes larger as shown by the data points marked by the symbol "+" when compared to the other data points with the identical total percent traveler shift. The influence of a dispersed program when compared to a more concen- trated effort is more clearly shown in Figure 35. Here, the curves repre- sent the trend of energy consumption versus percentage traveler shift for each successive ring of zones added to the program. As successive rings of zones were included in the work hour stagger, the general trend 112 .sowumesmcoo amuoom Hmuoa co mowoeaom Moo: xuoz manmwum> mo uommEH oedema Moor was: xmom eoum muoao>mua xuoz mo unwom useouom .em museum ow mm on MW mw mm .pm MN ON mH u 4 - sowusnwuumwo ammonium» Enemies moumowosH O mouoou newsman mcoam toumuucmocoo wowaom moumowosH < come soon: unease on medium...» hoeaom moumofioeH + mWirlI OH NH ma uoradmnsuoo Abxeua teqom u; aseezoea iueoiea 113 .mo:ON mcwummwowuumm emu mo GOHMMOOQ an coaumfismcou hmuwsm Hence no moeowaom use: xuoz manuaum> mo nommEH coaumm snow «as: xmwm as» 669m muoao>mua xuos mo amenm usoouem - u d d 1‘ cowusnwuumao ammomaou showed: moumoaosH AV mousou uwmcmuu msoam ooumuucoocoo hoeaom moumowodH 4 some means weapon on towammm mowaom mmumowoeH .LI v one .m .N .H mmcflm iomov H mafia N one H modem + $3 \m one .N .H more? a \V .mm magmas l OH NH va uoradmnsuoo Kbieua tenor u: aseeioaa iueozea 114 was for a greater reduction in energy use for a given percentage shift in travelers from the peak half hour. This difference became less pro- nounced as a larger percentage of travelers participate in the program. However, there was virtually no change in energy consumption with increased' participation in the staggered work hours when the program was concentrated in the CBD (ring 1). The anomaly of the relationship between staggered work hour partici- pation and energy use for the CBD can best be explained by the fact that the majority of the simulated work trips to these zones were relatively short in length (generally only to the 2nd or 3rd ring) and were routed over only a few highway links. Also, the highway links within the CBD were generally uncongested. The combination of short trips and uncongested links resulted in no change in the energy consumption. However, this result is consistent with the literature which suggests that the effect of a concentrated program on congestion is lost within approximately two miles of the program location. The variable work hour programs tested had a direct impact on high- way congestion except when the policies were concentrated in the CBD zones, as shown in Figure 36. The percentage reduction in congestion resulting from the variable work hour policies increases as the program became more dispersed and included more zones. The maximum decrease in the mean network congestion, based on the HCI, was approximately 44 percent. The relationship between highway congestion and energy consumption is shown in Figure 37. The relationship exhibits a strong linear tendency with an r2 (the square of the correlation coefficient) value of 0.97 for the equation 9 = 0.20 + 0.27x (16) 115 .soeummmsou amznmwm so moeoaaom Moon rho: manmeum> mo uoomEH mm. om me ‘ d 11 sowusneuumwo Hmuomeou anomaso mwumowocH mmusou uwmsmuu macaw ooumuucoocoo wowaom moumowocH some swamp euwuco ou ooeammm meadow moumowosH 605666 user name xmom scum muoao>mua Enos «0 umwnm ucoonom ON m o q iomoc a ones I? N one A mmowm 1T v 6:6 .m .m .H macaw m 6:6 .m .H mmcam .om museum OH ma ON mN om mm 0v mv om (IDH) xapuI uorqsefiuoa Kenubrn ueew‘paqurem eqa up aseezoea aueozea Percent Decrease in Total Energy Consumption 116 l + Indicates policy applied to entire urban area A Indicates policy concentrated 12¢ along transit routes 0 Indicates uniform temporal 1d distribution 9 = 0.20 + 0.27x r2 = 0.97 4 1 1 L 1 1 1 0 5 10 15 20 25 30 35 40 45 50 Percent Decrease in the weighted Mean Highway Congestion Index (HCI) Figure 37. Relationship between the Highway Congestion Index and Total Energy Consumption for the Variable WOrk Hour Policies. 117 where y = the estimate of the percent reduction total energy consumption, and x = the percent decrease in the weighted mean highway congestion index (HCI). The maximum reduction in energy consumption was approximately 12 percent for a reduction of 44 percent in the HCI. The reduction in the HCI would have resulted in an even greater de- crease in energy consumption had a mode shift not occurred as a result of the decrease in network congestion. This phenomena is shown in Figure 38 were the percentage change in work trip transit ridership is expres- sed as a function of the percentage reduction in the HCI for the peak period. The work trip bus ridership was reduced an average of 4 percent by an average reduction in the HCI of 22 percent. At this point, the elasticity of bus ridership to the highway congestion index is 0.18. The relationship between the percentage change in bus ridership and the percentage change in the HCI can be expressed by the equation: 3‘: = -0.65 + 0.212: (17) where 9 = the percent change in bus ridership, and x = the percent change in the weighted mean highway congestion index. This relationship was generated using simple linear regression with all the data points shown in Figure 38. The regression resulted in a r2 value of 0.91 indicating good linear correlation. This result indi- cates that a decrease in congestion due to the implementation of a stag- gered work hour program would have a negative impact on work trip bus ridership unless steps were taken to deter the mode shift. The possibility Percent Decrease in Work Trip Bus Ridership 14 12 10 118 -|- Indicates policy applied r to entire urban area A Indicates policy concentrated L along transit routes C) Indicates uniform temporal L distribution 1 9 = -0.65 + 0.21x r2 = 0.91 11 111 pg 1 1 1, 1 1 1 5 10 15 20 25 30 35 4O 45 50 Percent Decrease in the Weighted Mean Highway Congestion Index (HCI) Figure 38. Relationship between the change in the Highway Congestion Index and Bus Ridership Resulting from the Variable WOrk Hour Policies. 119 still exists that during an energy shortage transit ridership would in- crease even with the implementation of a staggered work hour program. Under conditions of normal fuel availability this does not appear likely. The impact of the staggered work hour programs on work trip length, time, and travel speed is described in part by Table 12 for both the auto- mobile and transit modes. The variable work hour policies had no apparent affect on transit trips, except perhaps for trip length. The data shows that the average transit trip length was reduced by 2.2 percent, indicat- ing that the reduction in congestion may have eliminated more of the longer transit trips. Automobile work trips were affected by the reduction in congestion resulting from the staggered work hours. The parameters most affected by the staggered work hour policies were automobile work trip time and speed. Figure 39 shows the relationship between the percent Of work traveler shift during the peak half hour period and the decrease in auto- mobile work trip travel time. The family of curves again suggests that concentrating these programs in a small area (the CBD) is less effective than a more dispersed approach. There is a distinct advantage in reduced work trip travel time through the implementation of variable work hour programs. The amount of the travel time decrease is dependent on both the location of the program and the number of participants. The decrease in highway congestion and the improvement in automobile work trip travel time are related as shown in Figure 40. As would be expected, the relationship shows nearly a perfect linear correlation with an I2 value of 0.996 for the data shown. The relationship can be expressed as: 9 = 0.217 + 0.370x (18) 120 Table 12. Impact of variable work Hour Programs on WOrk Trip Characteristics MODE: AUTO BUS Mean Standard Mean Standard Trip % Change Deviation % Change Deviation Length -0.23 1.68 -2.20 1.74 Time -8.73 6.08 . -1.48 1.10 Speed 7.87 6.13 -0.72 0.58 121 .mefi Ho>mua mans xuoz OHMQOEounn no mowowaom Moo: xuoz manoeno> no HommaH .mm ounmwh ooauoa use: was: ream Scum muoao>mua xuoz mo umwnm unmouom om mm om mv ov mm on mN ON m a L-llldlllllm .iJ. a inmob H mean N one H mmnwm nOflunnwuumwo Hmuomsou euomwnn moumowonH AV mounou uwmnmuu mnOHo . . . omumuunoonoo howaom moumowonH .< e onm m N H mmnfim mono nebun ouwuno ou oowammm meadow moumowonH m one .N .H mmnwm OH NH vH ma ma aura IaAezi dTIL 110M aIIqomounv U'EeW 811:} U"! 65981080 11.199185 Percent Decrease in Automobile Work Trip Travel Time 16 14 12 10 Figure 40. 122 '4- Indicates policy applied to entire urban area A Indicates policy concentrated GB b along transit routes 0 Indicates uniform temporal _ distribution ‘ 9 = 0.217 + 0.370x r2 = 0.996 1 L L 1 1 1 1 1 1 10 15 20 25 30 35 40 45 50 Percent Decrease in the Weighted Mean Highway Congestion Index (HCI) Relationship between Highway Congestion and Automobile work Trip Travel Time. 123 where percent decrease in automobile work trip travel time, and "<> II I: II percent decrease in the weighted mean highway congestion index. The decrease in automobile work trip time results in an energy savings as shown in Figure 41. Automobile work trip length showed only a slight decrease on the average. However, in some individual cases large changes were realized. This resulted from a redistribution of the traffic on the network (a change in the network assignment) resulting from the decreased congestion. 5.2.2 Group C: Staggering Werk Hours AlonggTransit Corridors Only one simulation run was performed with a staggered work hour program constructed to conform to the shape of the transit network while maintaining the base level of transit service. The results of the simu- lation are plotted in Figures 34 through 41 and are indicated by the symbol "A" for run number Cl. This singular policy was not particularly different from the other staggered work hour policies in terms of its effect on energy consumption, highway congestion, or transit ridership and was de- signed as a basis for comparison to similar policies that combined stag- gered work hours with bus redistribution measures. 5.3 Groups D and E: The Transit Alternatives Several attempts were made to reduce total energy consumption through a redistribution of the transit fleet during the peak period. The base distribution of vehicles was three per route (4.5 buses per hour). The basis for the redistribution was an evaluation of the highway network congestion patterns produced by MOD3. Initally an analysis was performed on the mean link congestion index (LCI). The LCI is defined as the ratio Percent Decrease in Total Energy Consumption 18 10' 124 Indicates policy applied to entire urban area A Indicates policy concentrated along transit routes Indicates uniform temporal distribution y = 0.02 + 0.74x r2 = 0.96 J 2 4 6 8 10 12 14 16 18 Percent Decrease in Automobile WOrk Trip Travel Time Figure 41. Relationship between Automobile WOrk Trip Travel Time and Total Energy Consumption. 125 of the actual link travel time to the free flow link travel time. The average LCI was computed for each transit route pair which taverse the same links in Opposite directions. Since the mean values showed a rela- tively small difference (approximately 4.0 percent) between route pairs, an alternative selection procedure was used. This procedure redistributed buses to route pairs with the highest LCI on individual links. For example, for run D2 route pairs 2-8, 3-9 and 6-12 (with maximum LCI on each route ranging from 6.000 to 8.999) were assigned additional buses while route pairs 1—7, 4-10 and 4-11 (with the maximum LCI on each route ranging from 4.000 to 5.999) were assigned fewer buses during the peak period. The overall policy structure for these runs is described in detail in Chapter Iv for runs Dl through D5. In each case, the base temporal distribution of work travel was used. The base distribution of congestion is shown as a histogram in Figure 42 for the peak half hour period with the links grouped according to their individual link congestion index. Only those links with an LCI greater than or equal to 1.5 are shown. Also shown in Figure 42 is the transit route that contains each link. This arrangement facilitates the tracing of changes in the congestion patterns as a result of the transit policies. The results of the redistribution of transit vehicles is shown as a network summary in Table 13. The link congestion analyses similar to Figure 42 are contained in Appendix D for these runs. The histograms in Appendix D for runs D2, D3 and D5 are the same because the travel and transit distribution during the peak half hour period were identical. The results indicate that although these policies were instrumental in reducing congestion on several of the most heavily congested links, the overall effect on network congestion was minimal, averaging only a 1.0 percent reduction in the mean highway network congestion index (HCI). kg 0850‘ I K 666:5: 6:66 a 26 .6660 when on» How nuouumm coaumomnoo H90: mam: room .Nv ousmwm “Bonn nodueeonoo x5e 11265 666.6 666.6 666.6 666.6 666.6 666.6 666.6 666.6 . 666.6 666.6 666.6 666.6 666.6 666.6 666.6 86.6 86.6 666.6 86.6 86.6 68.6 86.6. 68.6. 86.6 86.6 86.6 68.6 86.6 86.6. 66.....— 6 mad 6" 66 6 66 6 66 u an" 6 66 6 66 an «6 66 66 6 66 66 au 66 66 6 «on a an" 6 66 66 66 6 6n 6 66 6 an 6 66 66” n «6 «6 ea 66 «a 66 a 66 66 66— 66 66 66 6 66 66 66 66 66 oaa 66 n6 «6 6 66 ca 66 ~66 6 and a 66 a 66a 66 6 66 66a 66 as o- 6 66a 66” 66 66— 66a 6- 6 66a 6 66" «a 66. 666 66a 6 66— 6 66” 66A 666 u 66. «a 66— 66. 66" on 666 666 666 e 666 6 26 6 :6 6 z; 6 z; 6 26 6 :6 6 .Im; a as 6 :6 6 z; 6 (mm. 6 z; 6 as 6 26 a an 66 .63666 66:66 dance ODFIDIn'MNe-I OOFQnQMN—Ic Nun—16461616416616: ext-n 1mm!!! 30 rec—m 127 .Ammm3vmwn wuscwa mm.ma um mason “mm mmmsn my 650: Hmm 66653 0.6 663 mofl>umm mo 0.6: m.m m.ml 6.6: 0.6: 0.6 o.ml 6.0: 6.0: m.on m.o N.0I 0.0: 0.6 H.ml 6.0: 0.0: 0.6 N.~I 6.0: 6.0: 0.0 0.0 0.0 Nmumcm mmflue meCH mmwua Hmuoa 6603 mam coflummmcou x603 ous¢ xmzsmwm mmmu mmmm Scum mmcmno usmonmm 0 .v .m .N .H m 666606666 6666 xucmsvmum ammo 0663 was ¥ NH .0 m .m 6 6 .6 6.6 66. .6 66 .6 OH .6 m .m H 6 .6 0 .N 09 6 66 .m NH .0 m 06 .v m .m 6 .6 m .N vo 66 .0 N6 .0 OH .6 m .m H 6 .6 m .N mo 66 .0 NH .0 OH .6 m .m H 6 .6 m .N Na 06 .v m .m H 6 .6 m .N HQ «musom umm mmmsm mmmsm 663832 0660630 momsm mcwmoq mcfi>wmomm cam 00 .02 66666 musom 66666 musom mmflofiaom 9666668 00 mmmcm>6uommmm .MH @6369 128 These policies resulted in a very small reduction in automobile work trips ‘and increased bus ridership by only 0.90 percent on the average. Total energy consumption was reduced by less than one percent. Further evaluation of the link congestion patterns suggest at least one reason for the disappointing outcome of these policies. In most cases, the links that showed an increase in congestion were links on bus routes which lost buses to the redistribution. This indicates that transit rider- ship had decreased on those routes having a reduced frequency of service offsetting the increased ridership on the other routes. This hypothesis was tested by the addition of more buses to the routes that were most heavily congested without reducing the frequency on other routes. The effect on individual link congestion of adding one vehicle to each of six of the transit routes is shown in Figure D6 (see Appendix D). The effect on fuel consumption is shown in Table 13 for run El. The addition of these buses represents a 37.5 percent increase in bus fre- quency on the affected routes. This policy produced the best results of the transit-only policies. Bus work trips increased by 3.3 percent, while automobile work trips and network congestion were reduced by 4.4 percent and 2.9 percent respectively, and total energy consumption was reduced by 1.5 percent. This result indicated that the earlier redistribution policies failed to increase total transit ridership due to the lost rider- ship on those routes with fewer assigned bus trips. Overall, the transit policies realized considerably less improvement in energy consumption than the variable work hour programs. The modeling structure used to determine the work travel mode choice for this research was calibrated for an environment of "normal" energy availability. Under these conditions, it would be difficult to reduce energy consumption through improvements in transit frequency alone. However, 129 this does not rule out the possibility that a greater savings in fuel could be accomplished through transit improvements in an environment with restricted fuel availability. 5.4 GroupsiF, G and H: Combined Staggered work Hour and Transit Policies Several policy alternatives were tested in an effort to determine the combined impact of variable work hour programs and transit scheduling policies. These alternatives were in three distinct groups and included . policies which combined staggered work hours with transit redistribution, policies which combined corridor transit redistribution with the corridor application of staggered work hours, and a policy which combined the cor- ridor application for staggered work hours with the coordinated addition of buses. 5.4.1 Group F: Staggered Wbrk Hours with Transit Redistribution Six alternatives were tested to determine the combined impact of the staggered work hour programs and the transit redistribution policies. In each case a staggered work hour policy from Group A was selected such that 10 and 30 percent of the peak half hour travelers were staggered for each of rings 1, 2 and 3 of the urban area (runs A2-A3, A6-A7 and A10-A11). The congestion patterns for runs A2, A3 and A6 (see Appendix D) were similar to that of the base case and therefore the transit redis- tribution policy used in run D5 was also used with these staggered work hour policies. This formed the basis for runs F1, F2 and F3. These poli- cies are summarized in Table 14. The congestion patterns for runs A7, A10 and A11 (see Appendix D) indicated that slightly different transit redistributions would be more suitable for cases F4, F5 and F6. These distributions were based on the maximum LCI described earlier, and the policies are summarized in Table 14. 130 Table 14. Summary of Staggered Work Hour and Transit Redistribution Policies (Group F) Routes Routes Number of Run Receiving Losing Buses Shifted Number Buses Buses Per Route (D5)* 2, 8 1, 7 1 3, 9 4, 10 6, 12 5, 11 F1 2, 8 l, 7 l 3, 9 4, 10 6, 12 5, 11 F2 2, 8 l, 7 1 3, 9 4, 10 6, 12 5, 11 F3 2, 8 1, 7 l 3’ 9 4' lo 6, 12 5, 11 F4 3, 9 1, 2 1 4, 10 5, 11 12 6 F5 3, 9 1, 7 1 6, 12 5, 11 F6 3, 9 l, 7 1 6' 12 4, 10 * Run D5 is a Time 2, 3. 3, 3, 3: 3, 3, transit redistribution policy related Periods 4 to Percent Change in Travelers During Peak Half Hour Period ~13.2 -24.0 runs F1, F2 and F3. 131 The results of the runs for Group F are shown in Table 15 along with the related staggered work hour and transit policy runs. In general, the results indicate that the combined policies of transit redistribution and staggered work hours were no more effective at reducing total energy than the individual policies. Runs F1 and F2 produced combined results that were less effective than both of the related individual policy simu- lations. This could be due in part to the fact that both runs F1 and F2 concentrated the staggered work hour program in the CBD. The work trips originating in the CBD were of a relatively short length and there- fore contributed only a small portion of the total energy use. The transit policy appeared to have a negative impact because the decrease in ridership on the routes that lost buses was not counterbalanced by an increased in ridership on the routes that experienced an increase in frequency. Run F3 showed a greater impact on energy than its related staggered work hour policy run A6. However, it also showed a slightly lower over- all impact on energy than its related transit policy run D5. The results are similar for runs F4 and F6 where in each case there was a slightly lower energy impact than for the related staggered work hour policy. In only one other case (run F5) was the combined impact on energy consump- tion greater than the individual impact of the staggered work hour program. Table 16 shows the effect of these combined policies on the work trip characteristics of length, travel time, and speed along with the same information for the related variable work-hour-only or transit-only policy. A comparison of the combined policy runs to their related indi- vidual policy runs showed no improvement in these trip characteristics as a result of the combined policy. 132 Table 15. Effectiveness of Staggered Work Hours Combined with Transit Redistribution (Group F) Percent Change from.Base Case Travelers Highway . Run During Peak Auto work Congestion Bus work Total Number Half Hour Period Trips Index Trips Energy Fl - 1.2 0.0 0.2 0.6 -0.3 (A2)* - 1.2 0.0 0.6 0.0 -0.7 (D5)** 0.0 -0.1 — 2.0 1.6 -1.0 F2 - 3.6 -0.1 - 3.6 0.5 -0.4 (A6) - 3 6 -0.2 - 2.4 0.7 -0.5 F3 - 4.4 0.2 - 6.2 -0.8 -0.9 (A6) - 4.4 0.2 - 5.0 - .1 -0.1 F4 -13.2 0.9 -11.0 -3.6 -3.4 (A7) -13.2 0.9 -14.1 -3 9 -4 8 F5 - 8.0 0.6 -14.0 -2.5 -3.3 (A10) - 8.0 0.5 -14.5 -2.0 -2.6 F6 -24.0 1.5 -32.8 -6.9 -7.9 (All) -24.0 1.5 -33.1 -7.0 -8.2 * Number in parenthesis indicates related staggered work hour policy. ** Run D5 is a transit redistribution policy related to runs F1, F2 and F3. 133 Table 16. Impact of Variable work Hour and Transit Redistributions on werk Trip Characteristics (Group F) Percent Change from Base Case Mean Auto Work Trip Mean Transit work Trip Run ' Number Length Time §peed Length Time §peed F1 0 5 0.3 -0.2 -0.8 0.0 -0.8 (A2)* 0.0 - 0.6 0.4 0.2 0.1 0.1 (D5)** 0.3 - 0.7 3.4 —0.4 0.2 -0.1 F2 -0.1 - 0.9 0.6 -0.5 0.2 -0.5 (A3) 0.0 - 0.9 0.6 0.6 0.3 0 2 F3 0.1 - 2.0 1.5 -0.5 -0.2 -0.6 (A6) 0 6 - 1.7 1.9 0.1 -0.2 0.1 F4 -0.1 - 4.3 3.2 -1.2 -l.5 0.4 (A7) -5.8 - 5.4 3.4 -1.4 -l.l -0 5 F5 0.4 - 5.0 4.2 -2.1 -l.1 -0.8 (A10) 0.5 - 6.0 4.6 -1 1 -1.0 -0.4 F6 1.6 -12.2 11.6 -3.4 -2.1 -1.4 (All) 0.8 —12.3 11 5 -3 3 -2.2 -1.2 * Number in parenthesis indicates related staggered work hour policy. ** Run D5 is a transit redistribution policy related to runs F1, F2 and F3. 134 5.4.2 Group G: Corridor Staggered WOrk Hours and Bus Redistribution Policies Group G policies considered the combined effect of a staggered work hour policy oriented along the north-south and east-west transit corridors and a coordinated transit redistribution policy; In this group, travelers originating from zones along the east-west transit corridor were staggered from time period 3 to time period 2 (see Figure 32) and peak half hour travelers from zones along the north-south transit corridor were staggered to time period 4 (see Figure 33). The transit vehicles were redistributed to service the altered temporal travel distribution. These policies are summarized in Table 17. The results for these policy alternatives are shown in Table 18 along with the results for the related staggered work-hour-only alternative (run Cl). Here again, the combined policies showed no measurable improve- ment in bus ridership even when travel was shifted to match the increase in bus frequency. The decrease in ridership in the corridor with a lower frequency of service appeared to have offset any ridership increases in the increased service corridor. Table 19 reveals that the combined policy alternatives caused a neg- ligible improvement in the transit work trip travel time, while resulting in a slight improvement over the individual policy alternative for auto- mobile work trip travel time. 5.4.3 Group H: Corridor Staggered WOrk Hours and Bus Addition Policies To more fully examine the relationship between transit supply, stag- gered work hours, and energy consumption a combined policy of corridor staggered work hours and the addition of transit vehicles was tested. This policy employed the work trip stagger of run Cl and added 6 buses 135 Table 17. Summary of Corridor Staggered Werk Hour and Transit Redistribution Policies (Group G) Percent Change Routes Routes Number of in Travelers Run Receiving Losing Bus Shifted Time During Peak Number Buses Buses Per Route Period Half Hour Period (Cl)* -10.0 G1 1, 7 2, 8 2 2 -10.0 5, 11 3, 9 6, 12 4, 10 2, 8 1, 7 2 4 3, 9 5, 11 4, 10 6, 12 . G2 1, 7 2, 8 1 2 -10.0 5, 11 3, 9 6, 12 4, 10 2, 8 1, 7 1 4 3,_9 5, 11 4, 10 6, 12 * Run C1 is a corridor staggered work hour policy related to runs G1 and G2. 136 Table 18. Effectiveness of Corridor Staggered work Hour and Transit Redistribution Policies (Group G) Percent Change from Base Case Travelers During Peak Highway Run Half Hour Auto work Congestion Bus work Number _ Period Trips Index Trips (C1)* -10.0 0.6 -17.3 -2.6 G1 -10.0 0.7 ' -17.1 -2.9 G2 -10.0 0.6 -17.3 -2.3 * Run C1 is a corridor staggered work hour policy related to runs' G2. G1 and 137 Table 19. Impact of Corridor Variable work Hour and Transit Redistributions on WOrk Trip Characteristics (Group G) Percent Change from Base Case Mean Auto Wbrk Trip Mean Transit WOrk Trip Run Travel Travel Number ' Length Time Spged Length Time Speed (Cl)* 0.1 -5.2 3.6 -1.7 -1.1 -0.7 G1 0.2 -6.1 5.0 -0.9 -0.2 -0.4 G2 0.2 -6.3 4,8 -1.6 -0.9 -0.5 * Run Cl is the corridor staggered work hour policy related to runs 61 and G2. 138 during time periods 2 and 4 to the corridor containing the staggered work trips. The additional buses were shifted during the peak half hour (time period 3) to match the distribution used in run E1. In essence, six buses were added to the fleet during time periods 2, 3 and 4 and distributed such that they would appear to contribute the most benefit to energy re- duction. This policy is summarized in Table 20. The results for run Hl are shown in Table 21 along with the results of related staggered work hour policy run C1 and transit bus addition policy run E1. In this case, the transit supply policy did not change the total energy requirements from the related run Cl even though there was a slight improvement in transit ridership. When compared to run C1, the transit ridership in run H1 had fallen off sharply as a result of the stagger of travelers away from the peak half hour period. Under the conditions simulated, the decrease in transit ridership resulting from the staggered work hour program could not be offset by the simulated change in transit supply. The information in Table 22 suggests that the combined policy alter- native (run H1) showed only a slight improvement over the related transit alternative (run E1) in transit work trip travel time, and was less effec- tive in reducing automobile trip travel time. Overall, the combined policy showed more potential to reduce total energy consumption, based on the reduction in congestion and travel time, than the related staggered work hour alternative (run C1). However, the results were nearly identical. 5.5 Policy Effectiveness in the Alternative Fuel Environments The alternative fuel environments, as described in section 4.5, were used to determine if the fuel environment could affect the impact of the policies tested. This was accomplished by increasing the price of gasoline 139 Table 20. Summary of Corridor Staggered work Hour and Transit Addition Policy (Group H) Percent Change Routes Number of in Travelers Run Receiving Buses Added Time During Peak Number Buses Per Route Period(s) Half Hour Period (C1)* -10.0 (E1)** 2, 8 1 2, 3, 4 0.0 H1 1, 7 1 2 2, 8 1 3 -10.0 * Run C1 is the corridor staggered work hour policy related to run Hl. ** Run El is the corridor transit addition policy related to run H1. Table 21. Run Number (Cl)* (31):“: H1 * Run Cl 140 Effectiveness of Corridor Staggered WOrk Hour and Transit Addition Policy (Group H) Percent Change from Base Case Travelers During Peak Highway Half Hour Auto work Congestion (Bus WOrk Total Period Trips Index Trips Energy —1o.o 0.6 -17.3 -2.6 -4.2 0.0 -4.4 - 2.9 3.3 -1.5 -10.0 0.3 -18.3 -0.8 -4.1 is the corridor staggered work hour policy related to run H1. ** Run El is the corridor transit addition policy related to run H1. 141 Table 22. Impact of Corridor Variable Wbrk Hour and Transit Addition on work Trip Characteristics (Group H) Percent Change from Base Case Mean Auto work Trip Mean Transit Werk Trip Run Travel Travel Number LengEh Time Speed Length Time Speed (C1)* 0.1 - 5.2 3.6 -1.7 -1.1 -0.7 (E1)** 0.2 -14.8 11.7 1.2 -1.6 3.1 H1 0.2 - 6.7 5.5 -0.9 -2.7 . 4.3 * Run C1 is the corridor staggered work hour policy related to run H1. ** Run El is the transit addition policy related to run H1. 142 to $3.00 per gallon (Group I), and then, in addition, decreasing the auto- mobile/transit differential travel time coefficient by 50 percent in the mode choice relationship. Complete policy descriptions are found in Tables 10 and 11. Corridor staggered work hours, and a combined corridor stag- gered work hour and bus addition policy were tested using MOD3. The results are shown in Tables 23 and 24. The results indicated that the effectiveness of a staggered work hour policy, or the combined staggered work hour and bus addition policy, may be dependent on the existing transportation/fuel environment. The total fuel savings resulting from these policies was clearly greater in the en- vironment of higher fuel prices than in the base case. However, the fuel savings resulting from the policies tested was less in the higher cost/ limited availability environment than in the fuel environment of the base case. The additional energy savings resulting from the work trip stagger was only 2.9 percent and 1.0 percent for Group I and Group J respectively, while the energy savings amounted to 4.2 percent in the base fuel environ- ment (see run C1 Table 23). The addition of buses with the staggered work hour policy resulted in an energy savings of 3.0 percent and 0.8 percent for Group I and Group J respectively, while bus addition reduced energy consumption by 4.1 percent in the base energy environment (see run H1 Table 23). This indicated that the majority of the fuel savings was a result of the fuel environment and not the TSM policies tested. If large numbers of travelers were diverted to transit because of higher fuel costs or limited fuel availability, the TSM policies tested would appear to be of little additional benefit in reducing energy consumption. However, an important point remains to be considered when interpret- ing these results. That is, the analysis assumes that all of the addi- tional transit passengers could be carried by the transit system. The 143 Table 23. Summary of Policy Impacts in the Alternative Fuel Environments Percent Change from Base Case TSM Policy Travelers During Peak Auto Highway Bus % Increase Run Half Hour work Congestion WOrk Total in Energy Number Period Trips Index Trip Energy Savings (C1)* -10.0 0.6 -17.3 -2.6‘ - 4.2 4.2 11 0.0 —12.9 -21.5 60.0 - 5.2 12 -10.0 -12.1 -31.8 59.6 - 8.1 2.9 I3 -10.0 -12.6 -32.2 62.6 - 8.2 3.0 (H1)* -10.0 0.3 -18.3 '-0.8 - 4.1 4.1 J1 0.0 -15.8 -25.7 78.0 -10.1 J2 -10.0 -14.7 -32.0 72.7 -1l.l 1.0 J3 -10.0 -15.1 -32.3 74.7 -10.9 0.8 * Run C1 and H1 are the related corridor staggered work hour and combined transit addition runs with the base case fuel environment. 144 Table 24. Summary of Policy Impacts on Trip Characteristics in the Alternative Fuel Environments Percent Change from Base Case Nfigger Length gimp. §E§E§ Lengph T§m§_ §pggg Il -1.9 9.9 -13.1 41.4 18.8 19.1 I2 -l.5 6.8 -10.6 39.8 18.0 18.6 13 -1.7 6.0 -10.2 39.5 15.6 20.9 J1 -l.9 7.0 -10.9 47.7 22.6 20.6 J2 -1.9 5.2 - 9.8 46.6 21.9 20.5 J3 -1.6 5.1 - 9.2 45.7 19.1 22.6 Mean Auto work Trip Mean Bus WOrk Trip 145 existing modeling system was not capable of evaluating transit passenger demand and route capacity. Under the extreme conditions simulated, it is unlikely that the large increase in transit passenger demand could be accommodated by the existing transit system. Actualdemand would be detered by limited capacity. Under these conditions, spreading peak period travel demand would improve the effectiveness of transit supply. variable work hours may be a necessity under such conditions along with an increase in transit system capacity. The overall effect would be to divert fewer riders back to the automobile than was indicated by the results and hence the TSM policies would have a larger incremental effect than shown. CHAPTER VI Summary and Conclusions 6.1 The Policy Alternatives For the activity pattern simulated by this research it has been shown that variable work hour programs (staggered and flexible) could reduce network congestion and hence reduce automobile work trip energy consump- tion. The reduction in total energy consumption (automobile work trip plus daily transit) could be a maximum of approximately 12 percent with a uniform temporal distribution of work trips. A more realistic goal of a 4 percent reduction in total energy could be achieved with only a 10 percent shift in work travelers away from the peak period. However, the effectiveness of a variable work hour program was also dependent on the location of the program. It has been shown that con- centrating the program in a small area, such as the CBD, was less effec- tive (approximately 85 percent) in reducing energy consumption than a program involving the same number of travelers who work at locations that were evenly dispersed over the urban area. This result is consistent with other research efforts (Tannir (1977), Safavian and Mclean (1975)) which indicated that the effectiveness of a staggered work hour program was lost within approximately two miles of the work place. Under the conditions simulated, variable work hour programs have been shown to have a negative impact on work trip transit ridership. The decrease in congestion during the peak half hour period resulted in a proportional decrease in automobile travel time, which in turn resulted in a mode shift to the automobile. A 10 percent shift in travelers from the peak half hour has been shown to result in a 12 to 17 percent change in the highway congestion index, depending on the location of the variable 146 147 work hour program. This resulted in a 2 to 3 percent decrease in work trip bus ridership. The maximum decrease in transit ridership was approxi- mately 9 percent resulting from the uniform temporal distribution of work travel. Staggered work policies acted against transit incentives under conditions of "normal" fuel availability. Transit redistribution policies did not significantly reduce energy consumption. Increased transit ridership on routes given supply priority failed to offset ridership losses on routes sacrificing buses for transit redistribution. The addition of six buses to the transit fleet (one bus allocated to each of six routes yielding a transit frequency increase in service from 4.5 to 6.0 buses per hour) resulted in a modest 3.3 percent increase in transit ridership and a 1.5 percent reduction in total energy consunp- tion. However, when this policy was coordinated with a staggered work hour program, bus ridership decreased slightly compared to the base case, and energy consumption was not improved beyond that of the staggered work- hour-only alternative. The combined effects of variable work hours and transit scheduling did not generally improve the energy consumption beyond the level of the variable work-hour-only policy. The policies tested did not appear to coordinate well under the conditions simulated. The total energy consump- tion was influenced only by changes in the demand patterns. Increased transit ridership on routes given supply priority failed to offset losses on routes sacrificing buses for transit redistribution. One reason for the absence of response to changes in the supply parameters may be the design of the base case. The urban area, the travel patterns, and the structure of the transit and highway systems were all nearly symmetric. Although some transit routes did exhibit higher levels 148 of link congestion than others, there were no clearly dominant congestion patterns or transit demand patterns that could be influenced by a redis- tribution of the transit fleet. The conditions simulated resulted in a mean transit work trip travel time approximately three times longer than that for automobile work trips. This large differential between transit and automobile travel time would certainly result in only minor changes in mode choice unless major reduc- tions in transit travel time resulted from the supply changes. .This did not occur for the transit policies tested. The possibility exists that extraneous environmental changes could result in a mode shift to transit regardless of the travel time differ- ential. This appears to have been the case during the oil embargo of 1973-1974 when limited fuel availability caused a substantial, but tempo- rary, mode shift in many urban areas. Under a simulated condition of higher fuel price and limited availability, transit ridership was shown to increase dramatically. With this prevailing fuel environment, the transit supply and staggered work hour policies were found to be less effective in reducing total fuel consumption than they were in the base case fuel environment. A large fuel savings was evident as a result of the increased cost (direct and indirect) of obtaining gasoline. However, the large increase in transit ridership could only be accomodated through increased transit supply and a coordinated system of demand management such as a staggered work hour system. This would result in a greater reduction in fuel consumption from the policies tested than was indicated by this research. Limitations in the modeling technique prevented the proper evaluation of the policy impacts in the alternative fuel environ- ments. 149 6.2 Research Limitations D This research study had several limitations which may restrict the application of the results. The limitations stem primarily from the simu- lation technique and its scope of application. Transit trips were not assigned to routes, nor was route capacity a consideration in determining transit ridership. This becomes a factor when simulating a condition where unusually large numbers of transit trips are generated. The transit fuel consumption algorithm did not explicitly consider the number of transit stops per mile, nor the effect of highway congestion on transit speed. These considerations could alter the work trip mode choice, although the direction of this impact is unknown. Overall limitations of this research resulted from the scope of its application. A hypothetical urban structure was utilized for the simula- tion. The modeling system has not as yet been sufficiently tested on an actual urban area, and it remains to be determined whether or not the modeling system generates a reasonable picture of a real city. The shape and size of the area simulated may also have had an impact on the policy effectiveness. This possibility was not investigated by this research. Whether or not the policies tested would be more or less effective for a larger urban area, or in an area with a different spatial distribution of population and employment, is unknown. The relationship between the policy effectiveness and alternative transportation infrastructures also remains to be investigated. Changes in the highway network structure or the addition of expressways may alter policy effectiveness. This may also be true for alterations in the transit network structure, such as changes in route configuration or the addition of a rapid transit system. Planning for energy contingencies is a complex process. The evaluation 150 of many policy alternatives is necessary for each individual urban area. The results of this research indicated that staggered or flexible work hour programs could be a valuable tool in reducing work trip energy de- mand, and should be given strong consideration as an operationally inex- pensive method of reducing gasoline consumption. The high potential for energy savings through implementation of vari- able work hours indicated by this study suggests that further research be done to expand on these results. This should be done with the objec- tive of answering the questions raised by the limitations of the research, to further expand the modeling system, and to test other TSM policy alter- natives individually and in combination. APPENDIX A APPENDIX A Documentation of MOD3 Program Changes for this Research A.1 Program Changes and Documentation A complete program listing and user's manual for MOD3 are contained in the work by Peskin (1977). The documentation given here indicates the changes to MOD3 required to duplicate this research effort. All of the line numbers indicated with Fortran statements refer to the program listing given by Peskin (1977). Every effort has been made to insure that the documentation of the program revisions is consistent with the user's manual supplied by Peskin. Table A1 shows the revised input data sequence required to execute the modified version of MOD3. Only those variables added to the input sequence through modifications for this research will be detailed here. The description of the additional input variables is given in Table A2. The following additions to the Fortran coding for MOD3 should be made where indicated by the line numbers specified. (These statements allow the program to operate while surpressing the incremental layer of growth. This allows for "short-term" policy analysis. Fortran Statement Line No. Read 10, IRUN, BASERN, NOGROW 250 BE(J) - BASIC(J)*ESUB 9350 IF(NOGROW.EQ.YES) BE(J) = 0.0 9350.1 1250 CONTINUE 9350.2 IF(NOGROW.EQ.YES) Y(I) = 0.0 10440.1 IF(XPOP2.EQ.0.0) GO TO 1675 12700.1 GO TO 1679 . 12710.1 1675 CONTINUE 12710.2 PERC a 1.0 12710.3 1679 CONTINUE 12710.4 In addition to these changes the variable NOGROW must be added to COMMON block in the program MAIN and all subroutines. 151 152 Table A1. Revised Data Input Sequence Run Type if. m ea c 3.." cat. 3 c o w o 3383 m c s a Number ““7““ DESCRIPTION ‘ofCards X X X X IRUN, BASERN, NOGROW 1 X X X X KOMMNT 1 X X X X CODE, N, NARC, NITZ, IFIB, FLAG, NDUM, FLAG4 1 X X X VAR1, VAR2, VAR3, VARA, VARS 1 X X X VAR6, VAR7, VAR8, VAR9, VARIO 1 x x x VARll, VAR12, VAR13, VAR14, VARlS 1 X X X X WKINZN(30,5) 20. X X X X HMWRK 1 X X X X VALUE(5) 1 X X X X AUTO(40) 5 X X X X BTUPER 1 X X X X PRICE, OCC(5), OWNRSH 1 X X X X TRAFIK 1 X X x X I, J, AA(i), BB(i), CC(i), Z (Highway Network) NARC X X X X FIXD, FIXSP 1 X X X X PARK(N) N/8 X X X X PRKTM 1 X X X X WLKTM 1 X X X X PKRATE 1 X X X X NR 1 153 Table A1. (continued) 1 Run Type 4.) -H U) H c: :3 a m a .3 a: 2 o «a o $333 «a c: :3 a Number m H d < D E S C R I P T I 0 N of Cards X X X FR(NR) NR/8 X X X FRDAY(NR) NR/8 X X X TRANST(6) 1 x x x WALK(N) N/8 X X X WAITWT, WALKWT, WPLS 1 x x x RANDOM(N) N/8 X X X FARE, TFARE 1 X X X I, J, TIME, DIST, IRTE, Z (Transit Network) varies X X X F(200,5) 125 X X X X ICOST, BASE 1 X X X X ALPH 1 X X X X BASIC(N) N/8 X X X X INCR(6) 1 X X X X CAPX 1 X X X X BETA(2) 1 X X X X GAM(N,4) 4 x N/8 x x x LA(N) N/8 x x x x DEN(N) N/8 X X X X S(N,2) 2 x N/8 X X XXX FS(N,2) 2 x N/8 X X X X FSPOP(N) N/8 154 incremental run from the base run as a PUNCH file. (Table A1. (continued) Run Type 4.) -a m H c 3 >. m c: .4 £1 2 8 a o a: a 3 3 m. m c s o N r m H d 4 D E S C R I P T I O N Of Cards x x x x FG(N) N/8 X X X X G(N,3) 3 x N/8 X X X X XCOORD(N), YCOORD(N) N X X X F(200,5) 125 x x x LA(N) N/8 X X X BASESV(N,20) 2 x N/8 X X X BASEPO(N) N/8 x x x BASEBE(N) N/8 X X X BASERT(N) N/8 x x x B(N,N) (Base work trip matrix) N x N/8 X X X NOFACTS 1 * Note: These seven data variables are transferred to the 155 Table A2. Description of Additional Input variables for the Adapted version of MOD3. FACT(NOFACTS) (10F6.4) A vector of factors between 0 and 1 applied to the origin zone of the work trip matrix. Up to five different factors are possible, i.e., NOFACTS can be any value between 2 and 5. These factors are used to proportion the total work trips entering the highway network from a specified zone during a specified time element (used only for short- term policy testing of variable work hour programs). FACTl (F6.4) This variable is used when only a single factor is to be applied to all zones to proportion the number of work trips entering the highway network. This variable can be set at any value between 0 and 1. This variable is usually used to test variable work hour programs. IZONES(20, NOFACTS) (40I2) An array containing the zone numbers that each individual factor FACT(NOFACTS) will be applied to when proportioning the work trip for testing variable work hour programs (up to 20 zones per factor FACT). Specified only when NOFACTS is greater than one. NOFACTS (11) The number of factors to be applied to proportion the work trip matrix when testing variable work hour alternatives. Up to five factors can be specified, however, 0 must be specified if this Option is BEE to be used. NOFACTS should only be specified other than zero when NOGROW . EQ . YES (0). NOGROW (110) A Flag. When NOGROW . EQ . YES (0), a no-growth condition is indicated and all future growth is supressed in the incremental run; This allows for the short-term testing of policy alternatives. When NOGROW - EQ - NO (1). The program will generate growth in the incremental run as originally written. NOZONES(NOFACTS) A vector containing the number of zones each factor FACT(NOFACTS) will be applied to when proportioning the work trip matrix for the testing of variable work hour programs. Used only when NOFACTS is greater than one. ' 156 The program block in Table A3 should be inserted between lines 5870 and 5880. Subroutine ZONEWT (Table A4) should also be added to MOD3. These changes allow for the simulation of variable work hour programs. Prggram block from card 5870.01 to 5870.25: Factor WOrk Trip Matrix for simulation of variable Werk Hour Programs. Developed as an adapta- tion to the original MOD3 program. EXTERNALS USED: ZONEWT INPUT: B(N,N) work trip matrix from base run NOFACTS number of factors to be applied to work trip matrix FACT(NOFACTS) a vector of'the factors to be applied to the work _ trip matrix (NOFACTS > 1) FACTl a single factor to be applied to the work trip matrix (NOFACTS = 1) IZONES(20,NOFACTS) vector of zone numbers that each factor will be applied to OUTPUT: IBIG largest number of zones any one factor will be applied to B(N,N) factored work trip matrix ALGORITHM: This program.block factors the work trip matrix for the simulation of variable work hour programs. Up to five different factors can be applied to the work trip matrix to proportion the number of trips to be assigned to the transportion network. Each factor may be applied to as many as 20 zones, or a single factor can be applied to the entire work trip matrix (see Section A.2 Testing variable work Hour Programs). A schematic diagram of the altorithm is shown in Figure A1. 157 Program MAIN Changes for Simulating Staggered Work Hours Table A3. BER 0F #:créés T0 or APPLIED To HORK nnrnxx MUFACTS APPLIED TO WORK HATRIX o ) 0 p 0 1 E 0 u 1 I : N 0 ( G ( ITAT. N ‘L‘ ARIIPI“ AR‘R I AU T. D E o... L P P A! E B L L I u- ) as 0?: Tc Cnh AF F0 on. N 84H“. 0°C.... 11‘: CI on Q TTte) ET. OONI. GGCS 7:6. ‘0) N 01F;U 9007. QC 0 ‘1" APPLIEO TO U EERNqIIUN 0 0.5.1.2 a“ $58 010 TTMDCEDS To XPIG=NOZONES(I) {EU¢(IZONFSoFACT'IhIGQNOFACTS1 16) FF ITATRo-DZNZ 00“ ALAA=OOI 01h” [DHFH’MHNIDFFNLGIHTLTTxuvuo—u (N L N N CTOR TO BE APPLIED TD ENTIRE HATRIX .MR‘ 07—0 FOIFFOEOFOOBOFCA 900°: 0.? 5. K c 1010 Arum—P IF. 9T3... 1015 E R 1020 1025 I F 1030 1030 E R. 1050 ARFIF NI “ICCGCARFI 158 Subroutine ZONEWT Table A4. . 2 0 I ‘0 GI O 5 AN 2 O 5 L. 1 \I l\ 9 \l I F! R 0 OK) 0.3 I 9U CA 01R515 I 50 2V VIOLASSI) I RN 3 I 011591575 I I A 0 XISSCI OD.‘ .5 I I . Fc TIESCE)FOI Q I I QR ORV-l. Odeannns I I 8A 1..- 9T) 9(5.U.U.3 I C U“ fléVrlbrh‘uICTo-AUI I I s I x QXESSC OCIE I I EN TOXSISGIYO 9 I I O. QIOAPIQCOH’ I I CK SRIBAPIII 0‘. I T I N I. XAX 0C02T535 I NK I IN TVX) OP 9355 I I [R I SU I O ISISJN(n.5 I 9‘0: I II“ EQXSraFflJAOlu‘I I EHH IISOI HRX‘U O’DRRI‘ I F T. IST 0. IA 003).!TOPD I FE ITCE TV2P(0.C 009. 9 I IHND ICAD I 9.5-LB Ionic 035 I DTIN Inn—r0 E8XS$555X)5£ I A IFoc RR 0‘ QED—3 005V: I OPS ION O AAIB‘IIOIIOOO I U000 INCH FVS 955N955=Fu I II AT. I QTE ' .I QXISF ONE-‘5‘- I I 5.... IGCPU 07 05‘ 91ECC1 O \o I I elf—Rs I IA U0 .5 asp—«CIEJTX EC“ E I I SN R I BFTR "51:31.35“... 0 I...» .N T I AGED II 085 UV ILSEIT 0),.“ 0 c I .LZNT. I I) 00 s Oflauhn‘sa 0’05L 02 A I L If. ITSNN OhTEBGLISCSF ) To C. I CIA I CTR 0 ZRKS OF 0e 93 0 I 5 N I I TIUFN I ACEE MANILA) 9) 101.23...- ) .I R I \o I AFC 0 IFASB UVHBZ)6.§305F 5 C OI R I IREI I QFAI S O I QI:;I§.(C( 1 I In. T: T I MML..OHT IRTUrdF O: TIA 99‘." 996 (o F CIU c I IEUTA IEN OI 1RU5TOCSZ) IA .L 0 II I I PS c IMO-LO HAKSEON'NIIL N N F0. I. I 096. VI IOGS—D UVLIEndIII 9.3.? o O \0 Ir I U SLL IZIAI 3 'AL 1‘ QP..K550 z I YI N E I AYIPP IIBBF IQRADF15H5 or. 0 = L I O O I ELHPI I(I OI HR 98.: I41 OCSDRN I 0.1.. I N . I PPTAT ITCH 9 8A.: 05) 0L1U50C I SIPI )T I I I L IUSPT RVIIInISIe. O‘Cnn/ I TIA! Eon-A. I I LT .OU IEELS N 9R0... 95A...).u 915 I C1 S N QC 8 I NLYTH I NNAO U14AC$5ITC5 N 9.5.. ()ASOE 0.9.1“: I IUD. I 00 C JRV3 ofihLLKfi/RHNSTZFLTN 7.1L 0) I Tu. SE IZZII IA QIIILRLIIELORC IUN JIGSPIR I U XRH I IU I EVQAEPG OAHKS .2034 NOSZOequI :7 I OOIET IE H8 QC '15 OH IIUELAROTFF IZEII INTRC I RTRG I NNUUECIZR 95515 OTB—DLNCIJIONCJIOLTI I H THS IIO/STRRL150051 0/ II .n 0.....N0 90.....LUC 95.3.... I USAUM. III FAPAV5()5(5) RA F0(I=ZU(JIHIXUUU I SRuUNR IUSN 9R IV 04V5(F(3NELN 1T E 2?. : NNNN I 0 U I 0L0 QKH 93(35YOE IUG ON AHNN AOXHUDIIIR I STPFF I REMGPT IIG.....(&.UU5H.LLHIUH.DOIDHQERXuNTTTU I ICINH I ENHA OKRPTSNDNLSHTAH AR 2 AR 00 INNNIO I HQRUE I UIDLU?LAUAFRAA(ON—LOUEUOGDEOONFO‘OOOEN I TFTZP I SOC—PNFVVAbruFRVCCIRCARFDxmARFDIRCFCCCR—L I I 1923b5b789II E .L .L I t .R R F I I I I I I a 0 00° CC.CCP.CC.LC C 1 c 2 c 39..) 159 READ WORK TRIP MATRIX READ NOFACTS J NOFACTS OTHER ' _ LREAD FACTl l READ NOZONES ] L... . I G“ W) I {comm-.3 } Figure A1. Algorithm to Factor the Work Trip Matrix for Simulating Variable Work Hour Programs. FACTl 160 SUBROUTINE ZONEWT (card to ) Reads in at least two different factors and uses these to proportion the work trips originating in specific zones. CALLED FROM: MAIN (card ) EXTERNALS USED: none INPUT: IZONES(20,NOFACTS) vector of zones numbers that each factor will be applied to FACT(NOFACTS) vector of factors to be used IBIG largest number of zones any factors will be applied to NOFACTS number of factors to be used B(N,N) work trip matrix from base run OUTPUT: B(N,N) proportioned work trip matrix ALGORITHM: The work trip matrix is proportioned by multiplying each cell B(I,J) by the appropriate factors (FACT) for zone I. The algorithm is shown schematically in Figure A2. 161 Cm D 6 READ . FACT (NOFACTS) . I DO I = 1, V NOFACTS READ ' IZONES(I,J) B(I,J) = B(I,J)L *FACT I C m D Figure A2. Subroutine ZONEWT. 162 A.2 Testing Variable Wbrk Hour Programs The MOD3 program, as adapted, can be used to test the impact of var- iable work hour programs on traffic congestion, mode choice and energy consumption for the urban work trip. The variable work program adapta- tions include staggered and flexible work hour capabilities. As with other policy testing runs, testing variable work programs is done in the incremental run mode of MOD3. However, a no-growth condi- tion must be specified, that is, variable NOGROW must equal zero. The variable work hour option will only affect the total number of work trips simulated by MOD3. All other trip information computed will represent ‘ that-for an entire day's travel. In order to test variable work hour programs the length of the total time period to be simulated must be established. Usually, this will repre- sent the total PM peak period for the urban work trip. The peak period can be divided into any number of discrete time elements of any arbitrary length (15 - 30 minute time periods are recommended). At present, one simulation run must be made for each discrete time element in order to simulate the entire peak period. The proportion of the peak period work trips made during each indi- vidual time element must also be estimated. The estimate can represent the proportion of trips entering the traffic stream from all zones, in which case, a single factor (FACTl) is applied to all zones, or up to five different factors may be applied each to as many as twenty zones. This latter option will allow for a more accurate representation of the variation in work quitting times between zones. For example, Figure A3 compares three temporal distributions of work trips for a hypothetical 52 zone city during a 2.5 hour PM peak period. 163 Percent of work Trips Entering Network Numger - o Zones . uniform 20 Factors a) All 0 3:30 PM 6:00 PM 50 20 b) All 15. 1 10 L 3:30 4:00 4:30 5:00 5:30 6:00 50 20 15 G) 1-26 A 10 I 5 4* I fi 60 27-52 10 I 15 10 r——-J—5 I 3:30 4:00 4:30 5:00 5:30 6:00 Figure A3. Three Alternative Temporal Distributions of Wbrk Trip Travel for the PM Peak Period. * Note: The number of factors specified in (c) is 4 instead of 2 be- cause a single factor can be applied to only 20 zones maximum, and 26 zones are specified for each factor. Therefore, the input requirements would be that the same factor would be applied twice, once to zones 1-20, and once to zones 21-26. The factor would be applied in the same manner to zones 27-52. 164 The uniform distribution in Figure A3a represents the temporal dis- tribution of the network loading in MOD3 as the program was originally written. That is, the entire work trip origin-destination (O-D) matrix is assigned to the network for the time period specified by the free- flow link capacity. In this case, the link capacity would represent that capacity available for a 2% hour program. This is done through one execu- tion of the program. Figure A3b represents an example of how the temporal distribution of the network assignment can be changed to more accurately represent the peaking characteristics of urban work travel. In this case, five executions of the program are required (one for each half hour time element) to describe the entire 2} hour peak period. For each half hour period the factor indicated in Figure A3b would be applied to the work trip O-D matrix to determine the proportion of work trips to be loaded onto the network. In Figure A3c, two factors are used during each time period and ap- plied to separate sets of zones. This example represents a case where the work quitting times may differ between zones. As many as five dif- ferent factors can be used in this manner, each being applied to as many as twenty zones. Here again, five program executions are required to describe the entire 2% hour peak period. When the work trip matrix is proportioned in this manner, the MOD3 program output for work trips (including energy consumption) is reported as twice the calculated value for the time element being simulated. This represents both the AM and PM contribution of travel for the specified time element. All other statistics reported for the remaining trip types are for an entire day's travel. The work trip summary statistics for the entire peak period can be 165 computed by manually summing the weighted values from each time element. For example, the mean highway congestion index reported for each time element can be combined into a mean value for the entire peak period based on the number of highway trips occuring in each time element. That is n wti HCI = Z * hc. (A1) mean . 1 i=1 a where HCImean = the weighted mean highway congestion index for the entire peak period, n = the number of time elements the peak period is divided into, wti = the number of automobile work trips occuring in time period i, ‘ n WTa = the total number of automobile work trips = 2 wti, and i=i hci = the mean highway congestion index for time period i. This should be done for all summary statistics such as auto work trip length, time, speed, etc., where these values are substituted for the highway congestion index in the above relationship. The same proce- dure can be applied to the transit trip summaries where transit person work trips replace auto person trips. The specification of the length of the time period being investi- gated if provided through the highway link capacity (vehicles per element of time, usually per hour) input for each link, or the adjustment of the link capacity through the variation in the parameter TRAFIK, which is also supplied as input. For example, the link capacity may be input as vehicles per hour of green signal time on the network. The parameter TRAFIK is used to adjust the capacity of each link to represent the avail- able capacity on the network which provides 0.5 hours of green time per clock hour for a time period of 0.5 hours in length, when 90 percent of 166 the total capacity is available for work trips. Therefore TRAFIK = 0.5 * 0.5 * 0.9 = 0.225. The input capacity of each link is automatically adjusted by the factor TRAFIK. This eliminates the need to repeatedly change the input link capacity values when time periods of various lengths are to be simulated. The actual testing of variable work hour programs begins with the simulation of the existing conditions to form the basis for comparison. Various alternative temporal distributions of work travel can be hypo- thesized and simulated through changes in the factors applied to the zonal O-D matrix as described on the previous page. These alternatives can be compared to the existing condition to determine the potential for imr proving traffic flow. variable work hour programs can also be tested in combination with alternative transit strategies by varying transit route structure, cost, or bus frequency. It should be noted that work trip transit ridership is related to the input value of the parameter FR(NR), which is the fre- quency of bus service. The energy consumption of transit travel is come puted using FRDAY(NR), which is the daily total number of bus trips per bus route. When segmenting the total peak period into discrete time elements the automobile work trip energy consumption reported as output will be twice the value calculated for the time element. This represents both the AM and PM contributions to the energy consumption. The transit energy consumption represents that for the entire day. In order to perform policy tests with variable work hour programs, the data in Table A5 must be included in the data input after the input variable NOFACTS shown in Table A1. 167 Table A5. Additional Data Requirements for Testing Variable Work Hours 1. When NOFACTS - EQ - 1 add Run e .u ”.1 m .-+ c: 3 >. ’° a .4 $1 28 a a) m H o o g g; 45 Number mH¢I¢ DESCRIPTION OfCards X X X FACTl 1 2. When NOFACTS ° GT ° 1 (2 through 5) add x X X NOZONES (NOFACTS) 1 X X X FACT (NOFACTS) 1 'X X X IZONES (20, NOFACTS) NOFACTS APPENDIX B APPENDIX B Data Set for the Generation of the Peak Half Hour Travel for the Base Case 055 055 0835 055 08 5 020 020 ".20 ".20 020 100. 1.30 100 10.5., 100 I I I I I I I I I I I I I I I 088 088 088 089 O. 088 230 230 230 7.30 230 I00 100 100 100 100 III III III III. 00220022002200220 0 022064106.10$Q10;JQIO 60101000100010001000 IOOIIIIIIIIIIIIIIIII I I I 00.050005000500050 582200...01.510151015101510 I I 0.15102000200020002000 322200IIIIIIIIIIIIIIIII I I I 3 Don-90008.000803033U 07 «1.100080551086109610 8610 o I I 086102000200020009-6 00 I 24%200IIIIIIIIIIIIIIIII R - I I I E P r. I I 0000.".00000000006000. 072Q5900005720....72057205720 506057201ufl.ofl.30003000.3000 92.0I3n0IIIIIIIIIIIIIIIII E 0 In. I I I 0 .0 . 0 c u . L I IOOIIOOIIUOIIOIII 1T2176001009.?CCP2008200920 N513008209000400000004000 E IIUQOOIIIIIIIIIIIIIIIII I: 0.0 I I I 1.3 2.0 1.5 1.6 1.6 1.3 .L R C N. I 3 9288 1.. I I 06920400040004000400005575510 I2 03 466400IIIIIIIIIIIIIIIIII 7501001I A . . . o o o 0 C E S I a IIIIIIIIIIIIIIIIIIIIIIIIIIII. 5555:.555555555555! 5.2.35.1: 5:35... =. 153.332: 3.3-335.317.535.13491.1.1....33?v~5!dl..1:3,\ IIIIIIIIIIIIIIIIIIIIIIIIIIII. 000000006¢Cnenu0afl£0000300000000.; 0000000000 00000.003020030030R.u I 555.58.555.5‘«455...... .- 5.55... c .. a .a .. u. a 111.111.111.11....IIIIIIII.III1.II,Z21 11111111111111!1555=.5.— cs 5.25:6. 11.11.1111.xfi.‘.11.11.1222220.220nun-0...: 404404Q0h44Q094022222222 I I I I . IIIIIIIIIIIIIIIIIIIIIIIIIIII! III1111.11111111111111111 956.02200220.0AL26..OZZSOOQ-JQLI In59au0.9.n050710.1225279latte-112 brutal.) 0002069206°.20692069202 I I I O I 0002430553535550R.4550 351-5.0543131...” 0929507.00121527219501... 0230a 1.: 796 3. 145.5%.0R955. 5.630354—3. 51311:... .. = .. .. : = .. .. ..000000000000000000000000000300.0000n..0000003000035 0000000.... 00000000001.239557894015230557390123955739012365575.3012305675901230! 123 0557891111111111222222222233315333333Q40 b QQI Q. I 55555 55555665554 168 169 IIIIIIIIIIIOIIIOIIIIIIIIIIIIIIIIIIIIIOIIIIIIIIIIIIIII IIII...-IIIIIIIIIIIIIIIIIIIIIIIII.55555550055555555555500555555555555025555555555553055: 55.356.55.35555555555555555555555555533333333333333533333333333333333333333331...:3333351.3.3332 1.3-31:331..33333I:3.3333331.!«2..3tu333333333 IIIIIII.IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII00000000000000000000000000000000000000000900000000000m 00000000300000n.00000000000000000000000000000000000000000300000300000000000000030000.00..39a 00003000000000000000000000000000005555539000355555555000.uCU.._u..55555.—.0n..3n.335555555501909 a: 5.555....é 095555.355: 005.25.. 5555005555511111229.2221111111122229.211111111222222111111112...22231 1111111221111111122111111112211111 \ SfifiatSSQSfiSSESESSSERS:555§595aS:Efifi555$5555555§5fi=.:5E5:= 55a:==fia ::555=fifiSSfiSSSRfisssa5555::077777777777777777777777???7777777(7777777777777777777 JOOJOJOOLOSOOOCBIOOOJJ}J0030003903.888898893b8880658898886885688893onhRéBSéEdaéchBGaBuq .00.0....00...O0.0.0.0....00......100000000000000OO...OOOOIOOOIOOOOOOOOOOO0.00.0000...OI 111.111.111.111111111 11111111111111.1115 I , 887873245567659326576879801098709182930545432¢3506172a .31.?“6‘7687980o’0011221.34A45:..6‘.7789924‘.22111 1 1133330I11 1 1 9.2‘3130212 2 21224‘4‘229.12121.7 1313131313132325242¢242#242Q2I2422 23546567Q52356577889109 07890319263# 545231.‘ 0516271 266607585970809102132¢3Q4555677888 11 1 111333311 1 1 122343312 2 212?ZQI¢Q221212121 .011.11u1312u.&iu232365424““9.“ 2‘ 240‘. 2. a.“ as .. .. .. .. : z .. z .. z: .. ... .. .. .. .. .. .. : .. = .. .. .. .. .. .. .. .. .. .. .. : DUO-000000000000000000000000000000000000000000000000000 000a 00.000000000000003;.00000000000n.0123‘567890123456789012545678901234:.c.7 8.301230567890121. 6759012305578901234567890123Q567890000000000111111111122222222223333335333Q0QQQQQIQQSSER 56657777777 1.777888388889899999999991.111111111111111111111111111111111111.1111111111111111... 170 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII.IIIIOII0.0IIIOOIOIIIIIIIIIIIIOIIIIIIIIIIIIIIIIIIIIIIIII. 5555555000055555555000055555555000055555555000055555555:.5.—45555555959555—55555:.555.—.5..2.55:”: 1.33.333333.333.333.33333333333333333333333333333333333322222222342221111.11111.111111111111111.1a IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII. 330000000000000000000000000000300000000000000000U0000000000050000000008.000000000000003030 033090.000000000000000000000090000000000000000000000000 00.0300000000000000000900900330900! .I 5.... 559990099 n. 555: 090...... a CI 055559 .9..00900055.I50000n n 00559 R. 0. 8.0.9888.- P.80000000000000900C0 3C0 0.10 1..11122222222111122222222111122229.222111122222222111111111111110n 0.000.... n. I. 999300309r. I 500 U 111111111111111111111111.... 55‘ 09 090000000009 0n 0000000000000000900090000000000031 333333.555: 555535.355 :3558.6Ir58r.. .659. E3 50 7777777777777777777777777777777777777777777777777775555555523337?7733333333331222392312: C.U.OIIOIOOI...IIIOIIOIOI.IOOIIIIOOIIOOOIIOIOIIOOIIII5.a.33.3=u.—U.JIII011110.IIIOOOQ4“.5“44-:“qa.fi.u III IIIIIQII IIII IIIIIIIIIIII. 30a 2150 659 82110213209254 31076762176876576988710656551729 314123“19.345678901234557 890123 0:. 1323311 2244II221 1 1133402211 11333311 112244 111111111122222I 69012#556891201122390523401676712677856678978015656172a31452341123Q56780012305678001234: 2233311 2244I¢22 1 11133442211 11333311 11224. 111111111122222? .. .. .. .. .. .. .. : .. .. .. .. .. .. = .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. : .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. = .. .. .. .. .. .. : .. .. .. .. .. .. .. .. .. .. .. .. 0000000000000000000000000090000000000000000000000000000000000000000000000000000000000000 9567890123956739012305678901234567890123455789012305678901234567390123956739012.945.67.390... 55.555555565555557?777777773393888333999999999900000000.0011111111112222222222533333333300 1111111111111111111111111111111111111111111111222222222222222222222222222222222222222229 171 000000 5 o 7555:...) o . o . o . 500000 . 3. .5522! 000 ".00 ...... 44 7o.66o 0000 00. . 11 . .1 11991 000000 5 . 755555 . . . . . . 500000 . 3 .22252 000000 ...... 44 7666.6 000000 o . .1 o 99919 0.00000 '3 . 47.55555 . . o . . . 500000 . 3 . .22252 00.. n. 09. ...... 449765635 000000 . 1 o . .1 o R 49919 1 2 . 000000 R . 1755555 . o . o o o . . 1 500000 o 3 o .22522 000000 1 1 ...... 447765.66 000000 8 . o1 . . . 3 99199 .2 2 o o o . . o . . o . . o . . o I o o . o . o . o . . . 0000000 55 . 015755.355 o o . o o o . 111111.Q_1111919.229.I/.9.6.6_9..27v1.1 33! 55555:..5: :..555555=.,5:..5:15555555 0500000 o .33 .8 .55252 1000900 111111111111111111111111111 o o I . . o o 44440 .7 . .6 o . 1000000 33 11.1q R. 11911 4 2 o . . . . o o . . . o. . o . o o . o . o . o . o . . 0000000 55 . .7575555c.0 . o o . o o . 55573537555575535355?c.1557 3: 000009000000000000000000000 0500000 o .3‘ ..0 .22252 .1000000 077 .535 .7787 .0753570 .7107 .51 n.00000000000000000000000000 . o . o a o 0 44445 .7566 o .5 .88 5 ..3 ”CC“ .8 .023 .r3 6 . C4 .3 :J 000000000000000000000000000 1000000 03 . o .14. . 1 o o o . . .1 . 1 . o o .1 . o1 . . 0900930....000003000990800I6.0.0 3 99.911 1 111111111111111111111111111 10w 835.3989 558895559. 65594894513555.30000000 55 . .557555555 . . o . o . ...—3553918355559. 25319.22. 63425:... 1 1.3221....322333223332233327777 .050 0000 o .33 .5.. .22522 .10000 0001 o . .1 .1 o . .12 . . .1 .1 o . .23 . . .1 445544455444554445n.4445..08&80 o o o o o o . 44 449 .766 .A. .1 . .3322;.323.I .62437N21u1 4262432.. I o......oo..o.o..ooooo...ooo11000000 53 0.1.9 03 . o .3. o. o 3 99191 3 3 2 3 I.” 24. 3 5 4 678901234567390123455700 01280000000000255ooés7555555.......56765439n54356871416654798921 22293333333331.444 44444444 45510590000 . .51 . .33 .d .22522 .1000000311 1224.331 1244431 oooooooo21. 44447 .766.6 .1 01.000900 2 03 . .1 .9. 0 4 99191 6780. :12345678901 34557 769012 a 3676543905445687141655070 899 29.422333333333344 444.44 444555 8 311 1223331 1244531 .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. . 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000. 234567890123456789312345676901234567890123456789012345678901234567990123456789012345678c 4444444455555555555556665555777777777788538883839999999999000000093311111111112222222229 .2 2222 «229.229.2222 22229.222222222 2.9.222222222 2222.2 2222 29.222222 2333333333333333333333533333331 172 1. 1 1 1. 1. 1. 1 1 1.2.5.333Q‘ Q ‘ 4“ 4“ ‘.55:...5§.5p5‘255=,_rh6666656665777777 ’17 777q 82.0..9 9.5 O..G.O.Q.fi O 6,0 O_° O O 0.0 n.7CflsUGOP fl .-sthnalllll! 111111111111112o117 3w 55155573537555:.755 35355751557353? 551:. 5573531 555:.75535355751.357353755155573.“ 37... 5:. fi? 551...... 5 3101077 .535 .7757 93753570 .7107. 0535 2.131077 .535 .7707 00753579 .7107 .535 0701377 .52 .3 o? 707 .075! .3 5 .4 .88 5 a: .68 .3 .83 .58 o 85. ed .3 05 .w 0‘ .83 5 0J 83 08 .83 23.6 o .5“ o .c 5 05 .0 .4 0.6.6 .3 03 0.. t .8 o. ..3 o o 0101.. O 0.101.. 0.1 .01. 0 110.100 0 00101.. 0.1 0.10 01.100 0 0 0.10100 1 1 1 1. ...n a—SZQSSSPQ 1.9... (65.35589.Sfl;1352322‘258819R524:.559glfiupssssszq 3.13.59. “5‘25“...“ 1‘? 32224.“.52Nu92voh133558 5P2: 3.1.. .22.:ro 01.11.239.1NZ31 062412321021...“2624323232... 42 .332323231. 3524332334252“!236.12.21.14,0332323231.o6..?4.33.. .3 3 o o o o 0 SJ 5 o o o 0.5 o 0 1e 0 o ... 4 3 3 2 .3 I. 26. 3 3 .9 3 3 2 3 1“ 2n. 3 3 Q 3 3 2 ... 527728321327810801¢<439210925571410321‘509215676356613278G.547725§3980730.1871A12388012921 12452111 29.3331 24‘331 11245221 1134421 1333421 13... 22 111244t2 15627788321327 810901‘5“.9210Q 25671‘10323‘5092156u6‘. 5561327 81B52772£u§~3° 8978019v714123flno.0129?. 12‘421121 222331 n¢§‘§31 112*4221 :. 21. 1.2-3.52.21 112.322 111.2.5‘42 . 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 .2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 00000009000003000000.0000009900000.0000.90000000000000090000000.0000000000OOUOOOOOOOOOOOOOOQ 01234567890123.455730 fi.123§5678° 0.12345678901.43‘567801012395678.901sz56782 0.1254510’ 80. .9123Q567. 33333333334§Q9‘#444Q555555555555556665567777777777833383333399§.9?9999909300009091111111 3333333333333333333333333333333333333333333333335333353333333333335333‘§ 9 4 .Q Q Q Q ‘4‘ QC. ‘Q‘O U 173 ...... 222?22 333333 SQEQSS 29292? 8554aa ...... ...... cocoo- ago. 900090 . .0000. .00... 0900.0 00.... ...-o. 11111.1 1.11111. "0000.0 9000. 1 111111 a L 6999 3093....0 3111 2 11.1 311111 c 11.1 11.111. 221. o o o o o 0 222222 333333 555955 222227. 8.35444 to o o o o o o o o o o o o o o o o o o o o . 0.00.960 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 0 11.1111 11111.1 900000 060.3. 1 111111 866999 soaUneo. .3111 2 111 311111 6111 11111 2.6). o o o o o o 0 6.22222 333333 55555.3 222222 P.554AQ o o o o o o o o o o o o o o o o o o o o o . Olafisflvfl.h. o 0 o o o o o o o o o o I O o 0 o o o o o o o o o o o o o a a 11111.1 111?.11 30.1050 0540C 1 0 111111 845.996. 303-030 .311} 2 o 111 3111.11 5111 lfig’nll 22...) o o o o o o o 0 0. 229222 333333 555555 2229.22 F.55h44 o o g o o o o o o o o o o c o o o o o o o . 1 1 announunu 1 o o o o o o o o o 0 § 0 o o o o o o o o o o o o o o a o o 0 111111 1.11.111 DnUnUnunuU 3.08..” 1 0 111111 36690.3 30900.0 311.... a; C 111 311.111 6111. 11111 22? 1.111122232222222 o o o o o a .5 29.22222333333555555552222222585.3443 o o o o o o o o o u o o o o o o o o o o o o a o . 11111111311111.1111 09000001 0 o o o o o o o o o o o o o o o o a o o 0 o o o o u o o o o o o o o o1111111‘.1111110oocococc.d.u . 01 0 1111111 28.660 96 333 403314511 52 0 . 1111 33111115511 1111.1 27 3557§15573537551 o o o o o o .0 2?22222333333355.‘555522222225855443 o o o o o o o o o o o o o o o o o o o o o o o o . 570 .7107 .535 .701 09900002 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 01.111111111111100,00093990.” at o 9.4 .8 5 on 8 ... 01 0 11111.11 29660.95 3303053.... 1.1-. o .1 o .1 o o o .1 a 52 0 1111 ‘ 3311111561... 1 1 . . 11.1.11 9.? 95291542588188524 0 o o o o o o .5 522222223333333555555522222225655043ocoo.no...oooooooooocooo. 100.21.00.1ologo30 00. 003.02 Zoooooooooooooo .1 o 0 o o o o o n o 0 ¢ 0 o o o o o c o o o o 0111111111111110C00003090I. 3.349.624 1627a913n¢13§ 28 0 1.111111 «(86690.6 33:uU~.U.U..J.-.31. o o o o o .1 52 0 1111 33111116611 .3 24 3 3 4 11111 2? 5b1092301327652955 o o o o o o 00052222222331.3333555555522222225855“A3 o o . o o o o o o o o o o o o o o o o o o o o o . 13324211 13‘ 00600330312 0 o o o o o o o o o o o o a o o o o o o o o o o . o o o o a a o o o 0.1111111111111100000003000 291 o .1 1111111 2866996 3300000331.... 52 01 o 1111 33111115611 0 11111 2? .1561012301327652 1.3‘33Q211 13 : = z : = = = = = = : = : : : : : : = : = : = : : = = : = = = = = : = z = = : 3 : : = : .. .. : = = : z = = : : : = = = : = : = : : = : : : = : : = : 2 = 2 = = = : : = : : = : .. 0000000000000900000000000.009000003000000000000fl00000000900000.003090050"00.300000300000000 899124§5678Q 01239567899125 456789012345 673901234567890125456789 0125956789012345676901234: 1122222222223333333333§Q‘4“~54QQ55555555556666566666777777777753333833359999999999009000 4C#QQ4Q4Q4Q§QQQQ44QQQQQQQQ‘QQQ‘QQQfiOQQ‘QQQ‘QQQQQQQQQ444949§¢604Q4§§944Q‘4“4444‘4‘46555552 174 .0 0000.. 0.0.0.309 0.05.000 000... 0.0000 00 000000 a o o o o 0 222222 111111 11 055900 30000.". . 11 .5 22: .- 5. 33222 22 11222 c o o o o o o o 09906.0 0.30000 0 o o o o o o o o o o o 0.0 03.0000 0 o o u o 0 222.229. 111111 11 055000 300000 11 522555 53222 22 11222 no 00.... 90.0..fia8 030000 0.0... 0.0... 00 000000 c o o o o 0 222222 111111 11 055000 505000 11 522555 33222 22 11222 no 00000000990000n000 .0000. .00... 00 000000 c o o o o 0 229.222 111111 11 05-3000 300000 11 522555 53222 22 11222 cocoa...00900309.".1000000out...o.oooooo 0000000000 cocoooo52222226111111 11150—75000 03000001 1112522555 33222 222 11222 0.0.0....Cocngn.n.0clooooouo00.000.000.00. 0000300000 ......052222226111111 1115055000 03009301 1112522555 33222 222 11222 a o o o o o o o o 00900905 1....00000 o o o o o o o o o o o o o ....,5555555555555555555.555555%55:.555555555555: 5R 5.1.556 0000090000 oo.ooo.52222226111111227752225888852222222587777779522777772252?222325. 11150R.5000 7102.3 0 o c a o o o o o 0 0111.11 0 0 0999999 0 0 011911111 0 9 9.1999396 cc 1112522555 9 o 0 .65554333365 0 o o a 056.3 0 o o o o .345 o o o o o o .543 o o o o o .4 2.2:. 11222 6.466 66666 222222 6777776 211111 0.3000091 33222 ooooooooooODOOOOOIOOOSOOcoono...o.noo.55555555555555555555555555555555555555555555555575 0000000000 on.o.o052222226111111722788522258777785222222258722322785227777722587.9 1115055000 03000001 . 3773 a o o o o o o .9999 o 0 011111 . 0 09999999 . 0 .1111111 0 o .9 . 1112522555 33222 a o 0 033655593 0 o o .345 o o o o .543 o o o o o o .345 o o o o o o .563 o .6 222 . 11222 6446 2222 66666 2111112 6777776 2 = : : .. = : : : = : : : : = : = = : : = : = : : = : = = : : : : = = = : : : : = 2 = : : = : = : = = : = = = : : = = = .. = .. : = = : = : = : .. = = = .. : a a = .. = : .. = = = = .. 000000000000000000000000000900000000000.0000000000000000000000090009000000000000000000000 67 8901239567 890.123Q56789n 1234 567 8901254567 89912345678991.23fi567 890123“567896.123456789C123 0100111111111122222222223333333333Q.6QQQQQQQ5555555 556656666666777777777799999999999999 555555555555555555555555555555555555.5.55555555555555 555555555555555.555555555555555555555 175 2.635% 0280,951200000000000055093328482Q3000000000000890519.3712664 0030001000007700311469 n, 662407n2298650000000000008500546938681000000000000700973655115100000000000046029670510 391649.09 41646.... 300093000001‘ 0938761Q 95100000000000005507696154:..100030n.0.3000004 8053122000 046015084057 30o....000000000680351320O 0000000." 0000.0n OR 20779? 7060090000030.0000013002616000 00930807 Sc .30000.00000000008605090721000000000000000530. 51d02.100000000030000qu 5909 24 1059?. 15821712100 000000000000007203147500000000000000000930...; 3843.000000a 00000000019050000300. 12690020000000300000000002301.10200000000000000000.0490000100000000000000001..065030000009 oooooooOo-ooo9.0.0.9000...-cocoaooooooooooooooooco.ooooo.ooooooooooooooooooooooooo-oo. 51 0000." 0000900920 0063 00000900810 00000." On. 0000190 1 5388732698153000000000000573#3596070758000000000000287023#73626500000000000015439197851 602$ 109167021000000000000023.618236.64242000000000000775129659625200000000000055737099210 209.7474.609516000000000000973072275950000000000000023241628569.0090000000020045661867400 26949395992120000000000009320599318100000000000000787044370710000000000000035885805000 0262a52179100000000000000604585709000000000000000094102682800000000000000007727Ea20000 84G u9311§40000000000000006515074.800000000000000000610019500000000000000000252.510000On 2200220000000000000000000880111100000000000000300055010100000000003300000001403n000000 IO.0.0.0.0.....00OOOOOOOOOOOOOOOOOOOOOO....000.000.00.0001.000000000000000...OOOOOIOOOJ 53 00000000000071 00000000000071 00000030000072 6477.33572089202000000000095250922768580000000000000638480975788000000000000563136211“ A 509a56287766707000000000035.26309090.7400000000000007338287 89123000000000000049255611613. 119173523890401000000000001092988911000000....0.000000107003825490fl 000000300000925000;...750n. 642155085362000000000000085006316506000000.0000?»0009906626098000000000000000.53378202000 62361985591000r...000000000089079454.310000000400000000254"95590000000000000000000,06803000." 72.700530160000004.” 0000a 000086008392100000000000000008257 52° 11000000000000000004402000006. 44000004.90000000000000000020210020000000000000000008.01000100000000000000000059150000003 oooooo-ooooooocoo-00......ooooooooooooooooooooooooooooooooooooooooooooooooooooogo-ooo. 53 0 000000000001 0 00000000001 0 000000000033 1 1 1 024524rs 25947.4000000000000867484 203 b9050000000000006?7994455976500000000." 0000‘ P72530363 691365451899700000000000009970068307720000000000003527337056 8720000000000004716951542 n 5911605 63035300000000000009021.5138611000000000000098Q 73744580100000000000008814O 173.325 26n 30329.. 604500000000000094 cA.099068100000000000000604106o.‘50....000000000000001016...1.0a.Can 629369179932000000000000081241161770000000000000005‘40786“6600000000000000023273403302 33911245210000000000000009175140800000000000000000.633006020000000000000000040510000000 39000500000000000000000005911020100000000000000000121.10101009000000000000003b10030339;_ cacao-0.00000.00.000.00.00.0.00oooooooooooooooooooooooocoo...0000000000coco-000000.00. ..H7 0100000000” 95 0000000000027 0000000000093 1 1 3021316535812000000000000590420689320000000000000034Q686417 550000000000000037596 85779 a 16p.3740565565000000000000201185689734800000000000001095975969070000000000001464.2.5401.1 964439298380300000000000082311227120300000.0000000000221669652303000090000008691773...»5.0”... 401716“60114580000000000001137660922400000000n.000003628119320300000000.0000006129070530». 32751510685100000000000000632177591000000000000000783823797100000000000.000062426030003 1 62901432000000000000000090994.55330000000000090000378.32285200000000000000003143100nChm 01110931003000.00000000000216.000200000000000000000025500310000000300003000000840003 0301 00.00.000.000..0000...0.0000000000000000...oooooooocooouooooo0000.00...00.000.000.000. 53 1 00000000000082 00000000000081 00000000000081 7321042397176.000000000000599.7097615026000000000000603301855631100000000000046320194754 98¢5077Q1537300000000000085530377393930000000000000149066417743000000000000405204835A0 Q773003197175000000000000593102594932000000000000053030033135200000000000000151026560m. 162206183700200000000000088370177254000000000000008219036950300000” 000000004310045110L 6a360704761100000000000006521061072000000000000000611003915200000000900009044710911505 5690045001200000000000000.765005701000000000000000013100343100000000000000007730003.0007. 15100000000000000000000006110033200000000000000000009002010000000000000000058000003001 a...coo.0.00.ooooooooooooooooooooooo0.00....a...0.00....00.000.000.000.coco-0.00....o; 63 . 0 000000000000551 0 00000000000070 0 000000000000551 0 .727?90¢.8591181010000000000068503758769200000000000070R91155353930000.”0.000000824659509qc . .78030639569600000000000007603516118784.00000000000049531922664.3300000a 0000007791611404n 8 277506.0110093020000000.0002‘.95675310Q5000000000000004569194904503»...000030000014.9892069.? . 4834061076751000000000000821240023§60000000000000040629409o 5500000000.0000003254950910 6735030045530000000000000422540122400000000000000025468710330000000000000005922140000 134908151000000000000000006561108300000000000000003.553050720000000000000000517.“ 0100001 24390102000000000000.000002690030000000000000000000857009.00000000000000000000330000000r 0000.00.00...0.0.00.0.OOOOOOOOOOOIOOOOOOOOOO0.000...OOOOOOOOOOOOOOOOOOO00.00....0.0.0 54 101 0 000000000055 0 000000000074 0 0090000000061 1 _ 4509106915296200000000000000384653137629700000000000005591080830760000000000007261918691 0505917637180000000000000000700850696240000000000000001646953203000000000000005626653065 300187026562380000000000000050896983946000000000000001244241000500000000000000381598592J 401252031727800000000000000872322649900000000000000030129139100000000000000000241‘03605 08630199654nv00000nunu.00nu00007.76993896.5000000000000000.36708v 03555000000000000000506073000fi 017100316100000000000000001259021‘0000000000000000094800913000000000000000001326000000 02820020000000000000000000938101000000000000000000080810000000000000000000007460000001 cocoooooooooococooooooooooocoo.cocoooooooooocoo.cocooooooooooooooooooooooo.oooooooooo. 04 00000000000091 00000000000071 00000000000062 .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 4567890123Q3678901234:678901234567890123‘5678901234567890123‘567b901230567890123§567890 999999000000000 01111111111222222222233333333334A...“ 6.0.. Q 06555555555566665666367777777777 8. 5.5555566666666666666.066.66666666666666666665666666666666666666666666665666666666666666665 176 00000000000000.380a0604 42000000000000000649079.. $38368 1699: 2 346440 000000 009636 6:.4QL7 0.00000000000006509942982000000000000000553128 806465 610025 4....1324 000000 31756 7 P26 09: 00000000000000620534610n 000000000000000P48133 13.3522 441354 2 01635 000000 19 9330 16.4996 000000000000005205?63200000000000000000coo... 0.0.0. 00.... 0.0.00 00.... one... coo... 000000000000002206656 100000000000000000270289 115403 0040.28 074077 000000 790 231 11 000000000000003105011000000000000000000n 22a 23 a 66685 477654 71295: 066 11111 n.00000000000003600000000000000000000000 33124 222111 322111 223008 044 cocooooooooooooooocoo-00.000.000.000... . 3221 1 000000000000030 000000000000 1 000000000000001628565122200000000000000139547 726837 711690 525977 000000 407106 469649. 000000000000006647170432000000000000000040734 840532 551093 065618 000000 657972 70780.-. 000000000000003455696930000000000000000079904 474330 769612 427631 000000 145340 0.57191 000000000000004295050000000000000000000 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 00000000000000596477.300000000000000000156143 736502 445939 240906 055000 700213 111 1 000000000000007929100000000000000000000863842 866667 488657 292650 000 22111 000000000000008900000000000000000000000 142 222111 322111 385061 033 000.000...coo.ooooooooooooooooooooooooo 3331 1 1 00000000000062 .000000000000 000000000000002914541170400000000000000270697 981080 322594 655345 000000 450597 011664 000000000000009527432572000000000000000240631 346567 553862 121684 000000 614716 1%6093 000000000000002272711350000000000000000335204 000000000000000362315200000000000000000ooa... n.0000000000000334 212300000000000000000059296? 000000000000000757301000000000000000000011531 0000000009083 00000009000 93 5493 out... 59949.2 057924 000000 c o o o o 0 055000 000. 044900 .0 0... 700213 22111 297171 0 o o o o o 111 1 000000000000009610000000000000fl.00000000 11142 222111 322111 290991 033 .ooooooooooooooa...cocoooooooooooooooooo 323 1 1 0 000000000043 C 0000000000 1 00000000000000.5604941607100000000000000272783 025132 9359. 0 632311 000000 124023 91: 060 n000000000000060244B4492000000000000000622987 186679 611624 697138 000000 179963 760596 00000000000000“205963231000000000000000700174 904290 217590 663456 000000 278286 67496.0 000000000000000522050600000000000000000coo... c..... o..... 0.00.. 0000.0 0.00.. out... 0000000000000023463503n00"0000000000000756975 890893 734066 964317 055000 799123 411 1 n00000000000004162010000000000000000000733722 856568 477578 60A077 000 11111 3 n000000000000019110n0000000000000000000.11322 222111 322111 376500 033 .3 coco-0000000000....-o.ooooooooooooooooo 322111 000000000000009440491345500.0000000000009206088869662975392200 947702000000 0636161433476 7.. 000000000000007107518542000000000000000151806302420073276o 97096939900000001513217266.305a 0000003000000015510667800000000000000009505503393422863105662u057550000300305397:742509? 000000000000005484702260000000000000000oo00.0....ocoo-cocoooooooooooooocoooooo-oooooooo. 0000000000000019817213000000000000000006614365080313586727623517551500000087992127611 00000000000000280910100000000000000000028331367866765236775686669107066 11111 7 n 000000000000004.60030000000000000000000 33243.9222111332211161101903044 1 ooooooooocoo.oooooooooooooooooooooooooo 1 3221 1 1 00000000000071 000000000000 . 0000000000000030050142625000000000000002960853565310123738452576929000000002790551421819 00000000000000597408411600000000000000027201401835075005496640493290000000553121694‘5820 000000000000001589039140000000000000000131150103685447391441732200503000002248524a 66 £9.91 000000000000001382016210000000000000000coo.00.000000.00000000.000.000.000...coo...cocoo- 00000000000000:948092900000000000000000766545370471545930401749183350:50008700312 711 00000000000000708001000000000000000000027852366966858240985669378327000 22111.1 fl000000000000032....0000000000000000000000 2434222111332211163500903033 2 ooooooooooooono...0.0000000000000000... 1 334111 00000000000065 0 000000000000 00000500000000.02295052979200000000000000619674534735776463665592973600000004902127.».414.309 000000000000002564141689100000000000000439136975936570328468709.,9375000000045206437.093160 0.000000000000051.6318096200000000000000015942737304.190556960715151.5000000020:..,u594024u.n...ufi 000000000000003499030010000000000000000o.a...coco-oaoooaoooooooooooococo-cocooooooooooo. 0000000000000005257S0100000000000000000765492169051045629430545703950550008701122 11 1 _ 0000000000000029720300000000000.0000000028013276877678238.357666320147000 22111 000000000000004930000000000000000000000 114234222111332211161109193033 oonoooooocoo.no...ooooooooooooooooooooo 1 3331 1 0 0000000000641 0 0000000000 100000000000000874169533810000000000000569321609056208986879799144900000008584106591760? 00000000000000044776195365000000000000092767371436233706454470924260000000891179577599EQ 0000000000000000636204 6300000000000000094 69617664610553995983a 2172t0000000210468150 6690J 000000000000000411325570000000000000000000.00.000.00...one...coco...cocoa-0.0.0.0000... 0000000000000003584150000000000000000006927500881591076622664973283505500087000229 11 1 00000000000000090430000000000000000000027315286566478247847584366377000 221110 000000000000000130000000000000000000000 114234222111332211163797183033 1 0.00.00.00.000600.00.00.00...00.0.00... 1 ‘u2a‘1 1 00000000000072 : : 3 : : : : : : = : : z : : = = .. = = = = = : : = = : = = = = = .. : : : z .. z : : = = : 3 = : = z : : : : : = = : .. = : = .. : : .. = : = : : : = = : : = : = = = z .. .. = : :M— 000000000000000000000000000000000000000000000000000000000000000000000000000000 0000 00000 234 3678901234 36789012345678901234567890123456789012345676901234567890123 45.678901213456783 3888353099999999990000000000111111111117.2222 222 33333333334444444444555555n5n5666566666 4666565666666665667777777777777777777777777 777 77777777777777777777777777 7 7777777777. 00000000000 177 10.0691 312847 905161 360070 713900 9.4...005 589.130 4.60705 9.34816. 4880.60 3:.60P3 30306.... ’9 044971 392.508 6 76060 069 012 031310 522011 611180 721101 312010 528112 911100 266966 8920 252089 70 0 017 12:.300 420000 980500 600000 700040 £050.00 605000 601000 .2000 01 5.73310 9204 00.00. 0.0.00 .00... 00.0.. .0000. 00.... .00... 000... 0.0.00 00.... .00... .0000. 000. 111 01111 11111 36 93 8 14 9 4 4 5 a. 6 8 7 1 1 513 n. .....17 7 2 5 1 c 2 W 2 5 2 G 2 1 1 1 4 1 1 110946 880255 675691 369977 848009 607897 453990 404079 072074 19190.3 2524 014139 44 .99 076042 500612 201301 132111 279100 707111 714001 505210 896000 863.001 4300 99 so 4342a 4 786170 1 0 6 068999 590601 320100 550000 330000 101000 709.000 9.00001 722009 342504 1680 coo... o 5 2 2 00.... 0.0... .00... 0.0... ...... .00... 00.... .00... 0.0... .00... O... 1.770 .5 1.0’8 4111 1 1 11 7.4 4 2 28 1 32 307 0 5 1 9 5 2 2 137 9 665 n 1927 4 4 3 17 93 133 1 0 8 1 32 2 6 3 0 a. 3 1 2 31 1 1 1 1 14.7959 4.56453 9.82362 3‘ 0060 60.6199 902697 03690.J 906699 76309:: 9.32268 896. Pt 023407 6746 417484 0549—. C 990 929 690012 301100 689111 c 16001 704111 612100 775214 121700 4&5... 06 2.: c1 584090 27 9999 148691 240.000 790000. 208000 009000 300000 301001 927008 900901 728001 1057 .0000. 00.... 00.00. 0.0.00 00.... .00... .00... 0.00... P00... .00... 0.00.. 00.... O..- 4111 1 11 11 1 81 17 382 1 7 R. 9 1 2 1 536 .3 9.2 1 1 2415 4 7.1.10 7 58 85 5.... 1 5 4 5 2 3 5 2 a. 5 4 a. 1 1 5 2 1 611955 866367 1.07.764 1300 08 34 1919 698506 361851 392500 846909 43299 0 965081 P415763 569 .1 643299 705008 7 660 61 103111 169010 327101 216041 326100 347015 213110 501001 053167 5771 868969 047918 086E01 380000 096000 304000 909000 303000 765009 680000 620000 633617 9240 0.00.. 0.00.. 00.... 00.... 00.0.. 00.... 00.... 00.0... .00... 00.... .00... .0000. 000. 711 11 1 . 11 11 95 231 9 8 1 4 0 5 3 113 9 78 14 2t 5 1 3911 1 1E 11 7 0 2 8 1 2 8 19 3 3 9 1 1 728.935va 3421301124916637106069676819896446114400192059 0026 659430? 5291682390 97359579885: 5 90047991695631091738005628210211.123018802911112210119126.212033110201020,1231611007034.346.900.71— 926A1870152189224600812010000200100039060005800000400000090001503020090507003n72¢6650569 0.0000900...-cocooooooooooooooooooooo00000000000.ooooooooooooooooooooococo-000.00.00.00. 111 111 9 1111 9 3 . 0 1 72 2 19 83 9 90 0 3 6 4 4 2 9 7 3 5 1 4 5 1 3 4 3 6 3 6 4 2 a. 7 1 fix. 8700600019984052541275616063074198071403493937103917070080778501502810595309130935443821 379698027455903131064983001115693011010411006111107701201455001610912021191010246555095: 00470778086098637539500090000790500090090009403040010000053301418090r..00=..00.00.000022550094 0.00.0000...-on000.000......OOOOOOOOOOOOOOOOO0.0.00...000.00.00.00OCOOOOOOOOOOOOOOOOOOo. 111 5 11.1 77111 0 4 1385 4 9 1 10 0 75 7 .133 12 5 9 1 9 2131 12 s. 4 3 1 4 121 9 1 2 0 5 133 2 2 4 1 6 5 1 1 3 1 2 1 4 1 2 ~ Q. 9.0255600985061101“ 4 99485081361769998990..67 00758919260664051538870970369 6.40. 56191029357¢J7fl¢~3i~rx 36434 0' 88217 8.525067691111n504212001130709.521182:361112-30161622919051900610. 111.119.0110.. 0163‘ (01,00 . 9.1979986264998037: 9180 018031.0460108094n.0060»..900021613000050209003080109... 01000078619600..:.. 0 O 5 11 7 11 1711 1 865 13 1 8 74 2 17 1 655 2 5 6 9 O 7 2 9 6 2114 15% 4 . 7 22 5 1 3 S 15 522 6 5 2 6 R 9 1 1 . 1 2 5 1 2223759384804092603930131087932509450998899489001520656700477590214612221299712113492034 97A949318942803617939130321919710084672111232910197391711219000203512203450070755158091h 607398 501 o89180147717058000207000031760000008000007030600702010060700007650030322A 570310 .00.0000.00.00.0000000.00O...IOOOOOOOOOO..000O...00.000.000.000...00.0000000000000000600 7 11 1011 1 211 1 55 5 17 9 61 315 8 7 3 1 3 1 0 3 113 2 663 7 01 4 2 8 3 9 9 1 1 0 9 3 1 3 1 0 1 2 2 1 3 2 4 1 1 .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 0000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000 01234567390121.456769012345 b78901234.36789012345673901234567890123456789012345678901234567 77777777773383338938999n9999990000000000111111111122222222223333333333444444444455555556 77777777777777777777777 7777778888888838888888888988838888368883886885888888883888888888 178 571573 613035 1.60937 1D.3321 934850 378.920 897860 529787 790989. 7.040.90 4.35.00?» 900.130.? 6 3 ... 9.7 567016 630233 715565 112592 5084 90 507860 117060 a 7525.... 880925 301030 946256 04 843.2 1. 41 307682 090764 632425 990670 402820 929300 546740 3864 24 02." 817 01.4110 521.042 . 03379 7. o. 0.40.. .00... 000.00 00.... 0.0.0. .00... 00.000 0.000. .0004. 040... 00.0.0 00.0.. o 8 32 2 5 13 0 1 21 61 62123 142 6 c..? 8 1 91... 5 5 5 7 3 3... 1 9 74 2 5 26 117 9 0 4 6 4 8 4 11 4 1 5 5 3 91 5 9 b 1 1 1 1 .00 004216 878553 315141 313694 622251 463190 4.84.840 513036 287720 766230 209537 05‘ 612 0 09 532328 019533 955294 640658 488730 159960 190591 295577 41805.... 3.54233 770214 183559 ? .07 63344.5 74694.0 610619 028975 255692 705140 127584 17 4396 4.23740 2.23061 330399 409739 5 .0 .00... 0000.0 0.0000 000... 0.0... 0.0.0. .00... 0.00.. .00... 00.... 00.... .0000. o 3 2014 1 1 11 21134 422 6 377 1 31 3 265 5 129 0 1 672 8 47 12 7 46 8 7 3 1 1 4 1 2 3 6 8 3 0 1 5 0 5 1 1 4 1 1 1 1 1 98 948077 990239 71734? 188201 444152 993071 174712 969717 712411 6441... 5 791213 1% 3822 7. 26 01.77311 851318 067 65.: 9.55315 284067 015049 69962.5 07924.1 235188 527797 234847 4 0512:. 4 48 30378 6 051311 86624.... 366643 326216 111051 780126 094413 667511 199334 583656 0079...:5 7. 0. 0.0000 004.00 0.000. 000.0. 000.0. 0.0... .0000. 00.000 0004.. .0000. 00000. .0040. o 1 2929 66174 a 921.0 595 9 321 3 111 S 215 8 11.1 9.1 1 944 2 1.0 11 4 14 41 a, 1 11 4 1 1 3 6 7 0 1 4 0 5 2 1 3 5 . 1 1 1 1 2.0 4 29564 001320 81117 7 258961 073055 065029 4 84015 72 6055 1 c377 2 1090.38 20.49 ... 7 357507 5 8.3 936581 994130 56 9596 081225 570073 485063 776066 878077 387685 369419 670585 403141 9 :7 287555 670130 022641 191473 418011 221006 398042 346077 434504 252065 321467 140544 7 o. 0000.. 0.0... .00... 0.0... 00.... 000.00 00000. 000.00 0.0000 .0000. .0000. 00.... t 35461 7347 266 8 110 7 12 21 3 1 5 61 18 .33 4 017 171 a. 562 3 1825 P 2 6 1 1 1 4 3 5 9 4 3 9 6 1 3 5 3 1 .212412440710035091199.5507 3063121928714241.2502712400992290 27 17198646119994.6616931952Q 0094? a. 25206754010900508055930502022220504670060034904150.950057141001141806......0120737 47301.34 912C 333047050808990010765705089000404070640130241054211600683130124601009918606825710331539£ ooooooooooooooooooooo...ooooooooooooooooooooooooooooooooooooocoo.........ooooooo..ooo.o. .1 1 3194 881 2515 518 8 3 5 11 1 44 15 181 313 14 166 38 211 47 719 18 1 2 13 55 93 .62 8 5 12 7 132 213 1 05923665041.3010052981001247339614315531676385197661815719333841350,2747096364320: 85750.4 ea. 972869950103342060609205083022804675850657944069586909897750467:750996012510663601354110 714048.650502307040362806039844.904526700135310,0310749001769.40322173010408fi0510170000841.90 000.000.000.000...o0.0.0.000...OOOOOOOOOOOOIOOOOODOOOO000OOOOOOOOIOOOOOIOOOOOOOOOOOOOOOx .3. 776 130 556 927 312 3 4 748 11167 122 07 276 11 120 31 263 c 4 a 22 1... . 3 4 8 a M 3 6 1 . u 4 2 1 13 32 3 71 8 2 1 1 1 . 1 5421268359100000013659761459120345451.3820.399435312048589397873573054229999.43696065139770.3114 .592092742706451070479466012722202059250051494067801209652510299074027009801350720339006L 75306138010139.7080598019301406.903322560111133 07667560065423034109609 02.051059605307050: 3 oooooooo.coo.o.coo.cocoooooooooooooooooooocoo-cocooooooooocoo-cocoooooooooooooooooooo.o 5 1 222 3 7 5 8 415 7 1211 7 342612 111478 292 83 754 71 132 3 515 9 911 7 525 4 . 1 2 7 7 9 z. 2 4 1 3 1 1 6 1 3 11 a. 7 3 n. . 4 1 1 1 . 1 1 642400411029741903417413874593639763322096861367865636670642188033136605224180852589286 1800762728031848707173530123528081455702367590339373057R06704220670774072001706203152154 059041394702429180379365013973904250060290001064237309910710041024062304000560030159097 ooooooooooooooooooo.000.0.00.000.0.00.0...0.0000000000..00000000000000.0000...0000.0... 1 7 410.. 9 1 2 8 5 nu 2 19161 2363.3 17312 256 5 312 7 311 1 8 0 2 1 5 111 1 1 3 3 1 1 6 3 1 9 6 1 1 3 1 5 3 4 0 2 3 0 1 1 1 1 1 3...... ........ .... ............:.... .... ...... ......z: .. ........ ...... :8.......... ......:.. .. ...... .. .............. ................: ..................:.... ...... ...... 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000_ 890123456789012345678901234567890123456789012345678901234567390123456789012345678901234: 5555566656557777777777839833388399999999990930000000111111111122222.422223333333333 4. 44 4. 44 88889883888588883388538.6883an3888,3838888889999999999999999999999990.999999999999999 99999.1 179 200410 021860 711190 166510 G110a .940 76 000 0.09. 64 9407 141102 928902 240633 8604 89 .6607 10 4‘ 0731 50: 530 2 7310.0 610 47 0 6.750: 18758 #43203 398405 863604 440901 260077 410007 320055 420930 682750 714350 671130 $170: .97253 111506 371302 014104 122201 370310 1.30103 260512 c o o o o a o a o o o o o o o o o o o o o o o o o o o o . .000. 009000 00.... 0.00.. 00.000 0.00.. 00.000.00.00. 021 1 1.: 1 9.11» 17...: 151? 9770 2122 553 182 281 00. 03 3 1 9 1 3 f 1 1 2 1 2 2 3 4 2 4 427028 113909 688880 403080 61201 47405 55.3804 314407 545409 525502 460198 500325 972639 91340.5 0 740.80 7774 40 318010 1 6207 44201 7727 00 3015506 993504 549802 636309 280663 3414.10 252.811 6645.70 266140 16300.0 14105 .65408 111201 024104 048808 163507 242602 150321 111340 coo... coo-o. ooaooo coo... coco. coco. coo... ...-o. 0.00.. coo... coo-co oooooo coo... 5 1 2 1 113 31 59 082 8 232 2 5 0 3 8 0 1 7 6 2 9 7 3 4 2 1 a 1 5 3 1 3 2 1 3 2 3 4 2 2 2 2 065120 a 56290 715049 556070 69303 25801 956808 151500. 720 807 a 08984 824422 521198 0 05860 356331 701.450 237050 361030 6.... 009. 27907 227107 441! 01 21.6 808 3339.08 445644 360800 067320 5% 6300 6.81160 44700.0 228050 110 07 65602 149802 231003 463807 653202 793111 611150 723170 ...... coo... one... coo... co... ...-o one... 0.0... oooooo coo... coo... coo... coo... 22 3 14 21 61 65 02571912 292 39 08 184 11 514 1 2 3 4 1 1 1.. 7 4 1 1 2 0. 3 4 2 2 2 1 1 628390 851910 511088 20200.0. 0060?. 67779 902803 44000... 890690 323266 817980 113360 654350 098550 2.17010 677044 302040 20.2..0." 92126 0.24300. 340202 681201 440536 564690 219350 323280 287180 146070 021091 00.802." 0790;. 93402 742600 140701 211803 662311 274720 173240 186440 ...-o. one... 9.00.. can... no... 0.... 0.0... coo-co oooooo coo... coo... 00.... 0.0.00 16 29 59 71. 0. a 0104 446 02148 78 71 441 22 2 2 3 13 2 211 16 34 42 21 11 1 1 010.206081121900020010 0 06700 1007704 401760240044001050.8210308871832484004950?.008010701524000616307 06060000640000.03708.....00310... 249593050592100 06016071344 8.030480801115500 0123006072080003540 00018240000700.0000 200200 La. 0?. 3367010.203121070204803021170. .0222020449701025510001.8230 o o o o o o o o o o o o a o o o o o o o o o o o o o o o o o o o o 4 oooooooooocoococo-00.0.0.0...on.oooooooooouoooooooogoo 52 1.8 . 30 3 5 0. 37756849 9120 1 4 122 210 13 23 42 4 1 6 2 6 4 3 1 2 1 95. 8.9. "909.. .0: 80 810904331915010601261021— O. .4 .3776“. 5107004803168901-301. 95331118200019. 0. 600.40g. 42500040450006567. 0508241090657948502000010380. .22750. 6001257701.1749.006‘3517702762000600700020 630002676000400250130047592050262115.30060004900 91:. 5330.6807746443116220023940001517000066700008720000247.0 o o o o o o o o o o o o o o o o o o o o o o o o o o o . o o o 0 ca coo.coo...cooooooooooooooooooooooooooooooooooooooon... 312 9 9 4 5 9 2 1 9 1 3145 41 3 6 3 111 34 50. 14 22 42 21 0.» m u 2 .2 209 6410 90114 84716314 150548444083 01.. 7576661400404001119005336860472284007760504222070864506082970206516773483710030040.30306= 47351501487590670:07042744a05331080614005038060803810020~237 50152210.0314233020150602035 .758474091246101325000367512015010200343070012406004020ooooooo a...cocoa-choooooooooooooa; ooocooocooooooooc.0000a...0.00000009000000000009000.0.6 5 0 2 2 9 8 0 2 5 2 415 6 326 1 121 5 21 112 3 3 _ 1 4 1 2 3 1 2 3 3 4 1 2 3 1 2 1 1 0 6 1 1 1 4 1 41919580190570 91024 9980172... 590.. 3681. 9316541519440 047 430905555930585781029010502305040350798097 8754052119 90722154035204.003214 0665 82041629102297070673507086980006564060200906039040800126180104222025 015101044 20014 6 512484022221100676000068204001340600926020060605007010coo.a...cocooooooooococoon-coo... ...-0.0.0.0....coco-0.0.0.0..cocoa-000.000.090.00cocoa... 73 92 .51 5 2 3 5122849 246120 337 404 684 06 4 2 1 7 1 1 1 3 2 2 3 2 1 3 1 3 1 . .... ......z:......::....:.. .........................222...“ ...: ...... ...... ...... .... ...-:3...- :::: .... ...... : .... : : .... ...... = :388: .... 33...... .... =000000000000000000000000000000000 0000000000000000000000000000000000000000000000000000000123456789012345678901234567890123 67 8901234567 8901234 5678901234 5673901234567 89012345678900.00000000111111111122222222223331 4 4 44 5555555555666656666677771.777778598035559999999999000000000000000000000000000000000 99999 9999999999999999999999994.999999999999999 999999999111111111111111111111111111111111 180 26 15.3.2.2 579593 70690.9 826320. 454076 47 0 155950 357018 504078 705064 6.01.011 0056.45 6109.. 0 0899.70 408020 108044 2040.: 4 606021 9-08030 961447 631102 1.1.1907 084305 a 451.30 651009 1:. 0 433160 616040 407021 201095 106067 309048 128116 04 2302 017 005 28329.3 613070 831090 23 .0 .00... .0000. 0.0... 0.0... 00.... 00.... .00... .00... .00... 4.0.0. .0000. 0.0.0. 0. 2.51 1.— 35 31 914 12 14 521.211 822 91 9a. 4 1 2 2 2 4 1 4 3 1 1 2 3 4 3 ... 0 830079 047001 6280.71 236009 528964 951479 19.8991 7289.02 439705 480704 238772 9900.62 72 0 838010 343084 673091 311035 216064 687558 902903 691608 281507 794900 641091 702002 6? 0 227010 147001 02.50.25 005055 010054 032122 063207 064502 018202 155500 360.000 16209.5 00- O 00.... .00... .00... 0.00.. .00... .00... 0.0.0. 0.0... .00... .00... 00.0.0 00.... o. 120 89 42 941 42 76 3912042138132 40 1 3 2 1 4 1 2 2 1 2 2 4. 3 2 3 2 2 9 151084 026066 239129 637048 446027 708477 147806 910204 527 402 933246 470390 6 5901.5 17 0 689.045 £53099 6:39.09. 316044 476591 859273 613503 407105 321700 559899 195260 155020.. 2.... 0 336081 131012 022167 009052 030123 0.6830? 1.61704 364601 2823P1 652803 a 05390 233044 2: O 0.0.00 000.00 0.0... 00.... .00... 0.0... 0.0... 0000.0 .00... 0.000. 0.000. .00... O. 44 22 01 12 70.2211821163508 110 12 1. 2 4 3 1 2 5 3 3 2 3 3 2 3 7 739024 617000 577015 302440 327199 603796 9884 01 154600 180803 283949 513702 009.000 ch. 4 696032 913055 665015 264443 330206 953687 524103 712701 330201 438613.7149c1 664033 76 1 021093 027077 045036 030115 065402 072116 134207 222404 180900 372331 166010 025000 16 0 0.00.. 00.0.. 000.0. 00.00. .00... 00.... .0000. 00.... .00... .00... .00... .09... 0. 61 8 F. 1 5 4 3 9 1111 329 535 712 00 3 3 111 ... 4 2 1 2 4 2 2 1 1 1 3 2 3 3 33074029700756650713H 49.048697503860490 0050590110917180.09760205003103064 6102190003. 44.61601. 57088026502..80999040.245fl 050107 0049106004590130761084.013030710801044089.090 4024920001146 2709 50005086201612120312147027 8048021800204430150623029080903506030560 82.61.0302 125009421144 30 h 00.000000000000000OOOOOOOOOOOOOOOOOOOOOOO00.0.0.0.OIOOOOOOOOOOOOOOOOOO00.060400000000000 1 1 1 2 5 1. 2 51 12 36 5 01 1 98 1 3 25 3 45 8 9 9 9 2 3 1 1 41 33 1 12 2 12 4 1 3 250515458053418516319 49.09670170473051517330821850090184 6226090296218474000624182702.3522 31h 0.509782030599410 06080150216019.03340930772042061401006728140203047066.00000591217072120209 68002336.00125146024500301810450142003028308.50640024022616201060500174000002600301585990.... 00.000.00.00..00..0.0.0.000...O...00.0.0000...OOOOOOOOOOOOOOIOOOOOOOOO40.0.00a...coo...- 45 7 21 3 48 21 2.3 4 15 5 62 9 4 9 2 4 1 9 41 4 0 4 00 1 13 9,2 32 41 1 1 2 4 o. 2 4 2.4 0. .741680492145781616730370075032072501017510291150080012549005841100486946004694716555961n 53°“9-fi.59.9~0605° 02902.b80200310.05700100100780.03007‘.80‘v0037371903018900130fidq C.02.hu56°..079t 6 809.1. 1804b612806150420847002062701007140700439890063507004340400210220321fi303024391986550049u 00.000.00.00..0...O.000............OOOOOOOOOOO0.0.0.0.......OOOOOOOOOOOOOI0.000.000.0000 21 2 6 11 9 1 2 0 6 1 2 4 2 4 1 2 0 2 212 1 0 6 11 3 1 2 2 4 3 1 2 4 3 1 3 2 4 3 8011300701143050023109000080300604085015907409509990507282022279208939041646977139209005 4008350300749030072805003060000309037081606502540990677846032600105396070746207050300007 800227020064200002210900107000010603002110110130022004612500404080055307016544303660400L 0.0.0.0000...6.0.00.0.0.000000...00.000.000.000...OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOJ 9 112 38 50 75 02 61 9 3 5130912 .....9 1 3 3 3 1 2 3 4 2 1 2 3 3 3 _ = : : : : = = : : : = = = = : : = = : : : : = = = = : = : = : : : = = : = 2 = = = = : : : = : : : : : : : : .. : = : = = .. = : : = = .. = = : : = z = = = : : = : = = = = .. = : .9. 000000000 0000000000000000000000000000000000000000000.00000000000000000000000000000000000U 456789012345678901234567890123456789012345678901234567890123456789012345678901234567890 .55535544 4 4 4 44 4 4455:4555555566666666667777777777 888888888.899999999019000 00000001111111.1112 .000000000 0000 00000000000000000000000000000000000000.000000000000000111111111111111111111 .1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111 181 Qc.‘6 56,2‘. 11.30 0 I O O 80 82 5934 5‘23 0 o o a 73.01 5301 57.37 a o o o I. 4703 1031. 5102 o o o o 5404.3 QECSJ 21001 0 o o O o 9 11 24. F5299 601790 5430‘ O O 0 O 0 0‘ 39030 99035» 2500B 0 o o o 0 60289 80.1‘7 Q7053 . o o o o 5 O 1 = : = = : : : 0000000 23‘5673 2222222 11.11111 11.11111 APPENDIX C 182 mm>wumcumuH¢ >0HHom mo cowumfluommo vaHmqu U XHDzmmmm hOH. OOH. OHN. OOH. hOH. ON. ON. ON. ON. ON. OO OvIH th NhH. OOH. OON. OOH. NhH. ONH. ON. ON. ON. ONH. OO OVIH OHO VNH. OOH. OOm.. OOH. vNH. ONH. ON. Om. ON. ONH. Om OVIH OHd OH. vhH. NOV. VBH. OH. OH. ObH. Ov. ONH. OH. OH OvIH de OOH. OOH. OON. OOH. OOH. ON. ON. ON. ON. ON. OO ONIH MHd OOH. OOH. OOm. OOH. OOH. OhH. ON. ON. ON. OhH. OO ONIH NHO ONH. OOH. OOm. OOH. ONH. ONH. ON. Om. ON. ONH. Om ONIH HH< OH. ONH. OOv. ONH. OH. OH. OhH. Ov. ONH. OH. OH ONIH OHd vVH. NEH. OOm. NBH. ¢vH. ON. ON. ON. ON. ON. OO NHIH Od mMH. NhH. OOm. NhH. mmH. ONH. ON. ON. ON. OhH. OO NHIH Ofi HHH. NhH. wmfi. NEH. HHH. ONH. ON. Om. ON. ONH. Om NHIH hd OH. HOH. Ohv. HOH. OH. OH. ONH. Ov. ONH. OH. OH NHIH Od NHH. OOH. va. OOH. NHH. ON. ON. ON. ON. ON. OO @IH m4 OOH. OOH. Ohw. OOH. OOH. ObH. ON. ON. ON. ObH. OO le vfl mOH. OOH. NO¢. OOH. mOH. ONH. ON. Om. ON. ONH. Om le md OH. mmH. «Ow. mOH. OH. OH. ONH. Ov. OhH. OH. OH le NO ON. ON. ON. ON. ON. ON. ON. ON. ON. ON. OO and HO OH. MH. om. "HHMH. 0H. u: I: I: u: u- mmmm mmmm qq¢ H m v m m H m v m m H :oflummflofluumm om>ao>cH .oz ooflumm mafia vowumm «SHE unmoumm mmaou cam "mucow HH¢ How u©m>Ho>cH mmGON cowusaauumfla now :ofiusnauumflo Hmuomame Hmuomfiwa mGOHumHHUme cam coaumHsawm use: xuoz mHQMHum> amMImmmmw 183 mHnmoHHmmO uoz H Oz OH. OhH. Ov. OhH. OH. Ov. ONH. Ov. OhH. OH. 3 :5 OH. ON. Ov. OH. OH. OH NO .O0.0V.>V .Ov.Ov.vv .Om.hm.Om .Om.VM.ON .hN.ON.HN .ON.OH.OH mH.NH.O OH. OH. Ov. ON. OH. OH H0.0v.mv .Nv.Hv.ov .Om.mm.Nm .Hm.om.ON .VN.mN.NN .hH.OH.OH .VH.OH.O HO OOH. Orv. OOH. OZ HHO Om OOH. OOm. OOH. Oz QHO «m NhH. OOm. NhH. OZ HQO mm NhH. OOm. NhH. OZ HMO Nm OOH. va. OOH. OZ HHO Hm v m N v m N uoHuwm mafia "mason HHO How :oHuanHUmHo HmuomEoB onuwm mEHB "ou>Ho>:H mQGON How :oHuanuumHo HmuomEmB COHummHoHuHmm “smouwm ©m>Ho>cH mmGON AwwacHucoov .OZ cam .HU anma 184 umnenz wusom ouaom gown you use: um m mmflm meoHHom >HcouuHmcmua mo coHumHuomwa .onHumm mEHu 0>Hm HHm MOM HH can 0H .5 .O .v .H mmusou Eouu NH 6cm m .O .O .m .N mmusou mo nomm ou man wco .>H:o “:0: mes xmom ms» How HH can 0H .5 .O .v .H mmusou Eouw NH can 0 .O .O .m .N mmusou mo comm on was 0:0 .cho H50: OHMS xmmm map How oH cam h .v .H mwusou Eoum O can m .m .N mwusou on man mco 0mm0 mmmm MO ND HO H .No UHQMB 185 HQDEDZ UHSOM 0050“ SOON HON H502 H0“ mmmHHm .v can m .N mOOHuwm mEHu u0m NH cam m .O .O .m .N mmusou on use vcm m .N monuwm 05H» Mow HH flaw 0H .h .O .v .H mmusou ECHO NH can m .O .O .m .N mmusou ou wco MHm; xmmm now HH Ucm O mmusou Eoum NH can O mwusou on man mco ustm .OH cam .v .H mwuzou Scum m ucm O .m .N mmuzou Op mmmsn o3u uanm AomscHuqouv .Nu mHnma 186 umnasz ounom wusom noun uOM use: H0 mw>HumcuwuH< >UHHom uHmcduB can Haom xuoz anmHum> coaHnfiou mo coHumHHommn mwmsm OHHMO mvoHuwm 05H» How HH can O .O .N .H mwusou Eoum .NH can oH .O .v .m mwusou on man mao uqum .nO can Eoum coHuanuu . mm 5m 28m coflsfifima man uHmcmuB .O< can Scum :oHusaHHu .mo 55 ECHO :oHuanuumHo man uHmcmuB .m< :du Eoum :oHuanuu .OQ Gnu ECHO :oHuanuumHn man uHmcmufi .NO csu Eouw :oHudnHuu QEHB vm mm .mU andB 187 Hwnfisz wuaom 09.50% £00m .HOM H50: H0 womd—m uwHSm mmuw>mm .N UOHme mEHu OcHHsv 0H vac O .m .v .m .N Eouw NH cam HH .h .O .O .H mmuDOH on mmmsn N uanm .H0 :5“ mo GOHH m .N mooHumm mEHu now .OH mam n .v \H wwusou Eouw NH cam O .O .m mmusou ou man wco HMHnm .HHO Gnu Eoum COHu can m .N mvoHumm wEHu mom .HH 6:0 b .O .H mmusou Eouw NH wan O .O .m mmpsou on man wco uanm .hO cam Eouw coHanHHumHQ H0 Oh czm 635203 .8 28.9 188 ocHusu 0H wcm m .m .v .m .N mmuson on man H 66¢ .N uoHumm maHu mcHusn NH cam HH .5 .O .O .H mousou on man H OUO .HU :su mo :OHu HmuomEmB ustm wmum>wm .N OOHHmm mEHu OCHMSG 0H ucm O .O .v .m .N Eoum NH vac HH .5 .O .O .H mausou 0» man H uMHnm .HU Gnu mo :oHu HmuomEmB umnfisz wusom «EH9 cam ounom 50mm Now use: no mwmam AcmscHucoov .mu oHnma APPENDIX D 189 Hmnfidz musom UHmamua u m umnfinz xCHq u 2A .Ho cum How mmOHUGH :OHummOCOU H50: Ome xmmm .HQ musmHm xmvcH coHumomcou xcHA OOO.> OOv.h OO0.0 OO¢.O OO0.0 OOv.O OOO.¢ OO¢.¢ OOO.m OOv.m OOO.N OOv.N OOO.H . 00O.> 000.5 000.0 000.0 000.0 000.0 00O.v 000.¢ 00O.m 000.m 00O.N 000.N 00O.H NH mp O NO O OO HH Hb h Oh Om ON v Hv O Om HH 5N 5 HO NH ON O HOH O OO O OO v 5O mm H HO me O hm N 5O OO O OHH N MHH H HHH . O¢ mO OO O hv H Oh 5O OHH MO 0H OO OO NH Ov 0H 0O O Oh mOH .5 OO OO mNH m OO H hmH Ob h moH 0H OOH 0H HO hVH NO H «OH OOH OO NH NOH MO 0HH HO OOH O O0H ONH NO v HOH HH 0NH HOH OHH O VOH HNH OmH OOH «NH 5 ONH OMH m OOH OvH OOH OOH O hOH N OOH m zq m 2A m zq m zq m 2A m 2A m 2H m Zn m 2A m zu m zq m 2H m 2A mm. .ugnufim MMGHQ HMUOH. mcsm OOHHom OmuomHmm Mom mmoncH :0Hummmcou O xHozmmmO mmf‘kDI-HVMNHOO‘QFKOLDQ'MNH HHHHHHHHHH sxuxq Ieuozxequx go zaqmnu 190 Hmnazz musom amnesz xch ll of. 24 .OQ can ma .NO mcdm How mOOHUCH :oHummOGOU Hzom “Ham xmmm .NQ OHSOHO xmosH coHummmcoo xGHA mmv.N mam.O mmv.O mmm.O mmv.O mmm.v mmv.¢ mam.m mmv.m mam.N mmv.N mom.H coo.» com.O ooo.O oom.m ooo.m oom.v ooo.v oom.m coo.m oom.N ooo.N oom.H m OHH O OO O HoH O mm mm Ow H Hm ON mm HH NN N Hm OO O NO N mHH NH mp . HH HN mm m on O OH O hm OO O ON v mm N ON O NO oH OO O HO O NO NO an H HHH NHH N mm oH HO NH av H ON OH om HOH OHH Om N NO NO O Na ONH mmH m mm OHH mm ONH HOH NO m ONH O OOH OOH OO OOH HH oNH HO HNH . Om NNH N HOH m ONH H OOH . OMH ONH HH HMH N ONH O NMH NOH NNH OOH H NNH NH NOH OOH O «OH «OH NOH OOH OOH 2O m 2O m 2O m 2O m zu m zn m 24 m 24 m 2O m 2O m 2A m 2A O OO ":3ozm mxch Hmuoe H H sxuyq {euozzaqux go zaqmnn HOO‘QI‘OLHV‘MNHOO‘QFOLDV‘MNH NNHF‘IHHHHHI—l 191 Hmnfidz musom n m nmnfiaz xcHA u za .Oo cam H0O mmoHO:H coHummmcou Hsom OHOO xmwm .OQ OHSOHO mecH :oHummOGOU anA OO0.0 OOO.5 .OOO.5 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OOO.N OOO.N OOO.H 000.0 00O.5 000.5 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 00O.N 000.N OOO.H NH O5 O O0H HH H5 O OO OO 5 O5 OO O OO ON HH 5N O OO NH ON O 5O O OO OO HOH 0H OO O HO OO 5 HO 5O OO O OHH H HHH 5O H HO OH HO O 5O H O5 O O5 OHH OO OO OO HO O5 5HH OO 0HH O 5O 5 ONH O NO 5 O0H NH OO O ONH OO ONH N 5O H 5OH O O0H OOH NO NH NOH HNH OOH H OOH NH 0OH ONH OHH 0OH ONH OOH OOH NOH O OOH OOH 5OH OOH O OOH 0H OOH O OOH 5OH OOH OOH O 5OH m 2Q m 2A m zq m zu m Zn m 2A m 2A m 2Q m 2A m 2A m 2A m 24 m zn m 2A ON u552m mxcHO Hmuoa O\GDF‘\Olfl‘VCfiCVvO HHHHHHHHHH squq Ieuozzanu: go zaqmnu O N 192 Hmnsaz muaom u m Hwnssz xcHA n zq .Hm cam H0O mOUchH GOHpmoocoo usom OHOO xmmm .OQ OHDOHO xmvcH :oHummOQOU MGHA OOO.5 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OOO.N OOO.N OOO.H 000.5 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 00O.N 000.N 00O.H O OHH O .5O NH O5 OO 5 O5 OO OO OO ON O OO HH ON NH ON O OO OHH O OO O OO OO O HO O 5O OO 5 HO OO O HOH H HHH HH H5 H HO N 5O NH OO O 5O O O5 5HH 5 OO 0H OO OH HO O OO OO O5 O 5O OO 0HH OO 5O NO ONH OOH OO H O5 0H 0O OOH 5 O0H HO O NO H O0H OHH OO ONH ONH O O0H HOH O ONH HH 0NH 5OH OOH HNH O OOH 5 ONH NH NOH 0OH 0OH OOH O OOH H 5OH OOH HH OOH OOH OOH 0H OOH 5OH OOH a 2A m z: m 2H 24 m 2A m 2H m zu m 2A m 24 m Zn m 2A m 24 O5 “:3onm meHA Hmuos H H H OH OH OH OH 5H OH OH ON HN NHOO‘QI‘OIDQ'MNH sxutq Ieuozzequl go xaqmnu 193 .HO cam H0O OGOHOCH :oHummOnoo Hsom «HMO xmmm meCH :oHummmcou xcHH .Oa ousmHm umnasz munom umnssz xqu 2H OO0.0 OOO.5 OOO.5 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OOO.N OOO.N OOO.H OO0.0 OOO.5 OOO.5 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OOO.N OOO.N OOO.H O OHH O 5O NH O5 O HOH O OO ON OO H HO OO OO 5 HO HH 5N NH ON H HHH OO ONH OO 5O Om OO O 5O 5O HH H5 OO OOH HO N 5O NH OO O O5 5 O5 OH OO OHH 5O O OO OO O5 O OO 5 OO mOH OO OH HO H O5 OH OO N OHH O OO OOH OO NO HO O NO 5HH HOH OOH OOH OHH OO ONH 5 ONH O OOH 5OH O ONH HH ONH NH NOH OOH HNH O OOH H 5OH ONH OOH OmH OOH OOH OOH NOH OOH OH OOH 5OH OOH OOH O 5OH m 2A m 2H m 2A m 2A m 2Q m 2H m 24 24 m 2H m an m 24 m 2A m 2H m ZR Cm «EOEW WMCHQ HMHOH. ommhmmfl'MNHomml‘Qmfl'MNv-i NHHHHHHHHHH squq teuozxanul go xaqmnu .194 Hmnssz musom u m umnasz xCHA u 2A .Nw new Ho .Ho mcsm mom mOUHOGH COHummOCOU Hsom OHOO xmmm .OQ OHSOHO meGH cOHummOcou xcHA OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OOO.N OOO.N OOO.H 000.0 000.0 000.0 000.0 000.0 000.0 000.0 00O.N 000.N 00O.H O 5O O HoH NH O5 5HH O OO OO. OO O HO ON 5 HO O OHH O OO HH H5 OO H HO HH 5N O 5O O OO 5 O5 OO OO O OO N 5O H HHH 5 OO 0H OO OH HO OO O OO OHH O 5O O0H O 5O 5O 0HH OOH NH OO O O5 ONH OO O5 ONH OO 0H 0O HOH H O5 O NO OOH HO OO NO OO OO HH 0NH H O0H HNH OHH ONH OOH 5 ONH O OOH O ONH 5OH OOH NH NOH H 5OH O OOH OOH OOH 0H OOH 0OH m 2A m zn m 2A 2H m zq m an m zq m 2A m 24 m 2A H5 "58% mxcHO Hmuoa H H H OH OH OH OH 5H OH OH ON HN NN NHOO‘CDI‘OIDVMNH squq {Buozxaqul go Jaqmnu 195 Honssz musom u m “Onasz xch u zq .Nm cam MOO mmoHOGH COHquOGOU Mao: OHOO .50 muanm xmocH COHummOGOU xch OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OOO.m OO0.0 OOO.N OOO.N OOO.H OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OOO.N OOO.N OOO.H H m5 O 5O O OO HH H5 H HHH Om ON Om 5N 5 HO NH ON m OHH O HOH O OO 5 O5 5HH OO OO 5O O 5O OO OO N OHH O OO H HO OO Om NH OO O OO 5O 5 OO OH OO HO N 5O OH HO O O5 O 5O OHH HO mO H O5 O5 ONH OO OO HH ONH OH OO HOH mOH NO ONH O NO OOH 5OH OO ONH OO OOH H OOH ONH O OOH OHH OOH HNH m OOH ONH OOH OOH NH NOH OOH OOH H 5OH O OOH OH OOO OOH O OOH OOH O HOH OOH O 5OH . N OO O 2O m zg m 2H m zu m 2; m zn m ..zg 2H m zu m zq m zu Hm «gogm mxdfld HMUOH. NHOO‘QI‘OMVMNHOO‘QFKDUWVMNH NNNHHHHHI—IHHHH sxutq teuozzaqux ;o xaqmnn 196 mm¢.h OOO.5 mmm.m oom.® .Om cam MOM mmoHocH coHumomcoo Mao: HHmm xmmm xmccH :oHumomcou xcHH mmv.m mmm.m mm¢.m mm¢.v 000.0 oom.m ooo.m oom.v ooo.v oom.m mmv.v mam.m .OO OHOOHO mmv.m ooo.m Hmnsnz wusom mmm.m oom.N OOO.N OOO.N umnssz xch a“ mom.H oom.H m mHH m hm mm Mb mm mv mm m m¢ m mm v mm H Hm hv m HOH mm hm HH Hh H HHH or mmH h mm mmH mm mmH N MHH 5HH mm l‘ HH m v OH OH O hm mm mm Hv Ho mo Hm hm VOH OHH mNH NH m mv mv mm no mo MOH mMH HvH th NH Hm hm no om mHH VNH NmH mmH 00H OQH NH H OH mm mm mh mm mh mm om mm mm mOH ONH HNH mNH mNH OMH MMH hMH mMH mvH mvH mVH me bmH 00H m mm 24 m 2H m 2H m 2H m “Ozonm mxcHH HOuoe 2H 2H ZH 24 H H HOO‘QI‘LDIDVMNH MH OH mH 0H NH mH mH ON HN mm mm vm sxuxq IeuozzaquI go zaqmnu 197 Hmnsnz musom u m umnssz xch n 2H .Om cam now OOOHocH :oHummOcou H50: MHmm xmmm .OQ mHnOHm xmvcH onumwmcoo xcHH OO0.0 OOO.m OOO.m OO0.0 OO0.0 OO0.0 OO0.0 OOO.N OOO.N OOO.H coo.O oom.m ooo.m oom.O ooo.O oom.O ooo.O OOO.N OOO.N OOO.H mm mm NH O5 OO O OO ON HH 5N OO 5 HO NH ON O OHH HO O 5O OO O HO 5O OO H m5 5O OO O 5O OO OO OH OO O O5 OO O OO NH OO N 5m OO HO NO 5HH 5 O5 OHH OH HO OHH O 5O O NO O .OO ONH HH H5 5 OOH OO O OO OOH O OO ONH O O0H O HOH O0H mOH OHH HOH H HHH HH 0NH OOH O ONH oOH OOH H 5OH O OOH 5OH OOH NH NOH N OOH 0OH O OOH m 2H m 2H m zq zq m 2H m 24 m 2H 2H m ZH m 24 mO "csosm OOOHO Hmuoa OO‘QFOMVMNH sxutq Ieuozzaqux go xaqmnu 198 umnfidz Ousom u m umnssz xcHH u zq .OO cam uow OOOHOOH GOHpmmOGOU Macs MHOO xmmm .OHQ OHDOHO xOOCH :oHummOCOU xGHH OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OO0.0 OOO.N OOO.N OOO.H oom.O ooo.O 000.0 ooo.O 000.0 ooo.O oom.O ooo.O 00O.N. OOO.N 00O.H OHH NH O5 HH H5 5 O5 OO ON O HO HH 5N 5 HO O 5O O OO 5HH O OO OO O 5O OO NH OO O OO H HHH OO OO O 5O 5O O HOH OHH H HO OO O OO H O5 oH OO 5 OoH OO O O5 5 OO 5OH N 5O O5 O 5O OOH O OO OH 0O OHH OH HO O NO ONH OO OO ONH OO OO HOH HO H O0H OOH NO O O0H OHH HH 0NH 5 ONH HNH O OOH O ONH NH NOH OOH 0OH OOH H 5OH OOH OOH oH OOH OOH 5OH OOH O OOH m 24 m 2A m ZH m 24 m 2H m 21H m ZH m 21H m 2H m ZH m 2H O5 “585 mxcHO Hmuoa OO‘CDI‘OLGVMNH sxurq Ieuozxanux go Jaqmnn 199 Total Links Shown: 54 5-hahahahahahahak>k>k>kakaN:k:k>kfk>u: N)h)erIG\\JG’E’CDF‘N)&’D|U‘O\\JGDWDCD Number of Interzonal Links ...; H ...; HNWume‘flm‘DC LN R LN R LN R LN R LN R 164 6 160 159 152 12 147 145 139 3 135 129 123 104 93 82 81 79 78 8 68 67 165 61 10 153 57 2 141 55 110 51 1 103 7 47 97 4 45 95 43 89 4 117 41 4 83 113 2 39 9 69 10 111 1 115 3 37 8 S3 76 7 101 6 33 49 12 73 12 99 5 25 35 71 11 85 8 87 9 1.500 2.000 2.500 3.000 3.500 1.999 2.499 2.999 3.499 3.999 Link Congestion Index Figure D11. Peak Half Hour Congestion Indices for Run F6. Link Number Route Number LN 56 II II Number of Interzonal Links 200 Total Links Shown: 71 LNRLNRLNRLNRLNRLNRLNRLNRLNR 29 166 28 159 27 157 26 152 12 25 149 10 24 148 23 145 22 139 3 21 137 1 20 130 19 128 9 18 124 17 120 11 16 118 15 96 14 93 164 6 13 92 5 160 12 82 147 11 79 135 10 78 8 123 9 75 l 104 1 165 67 95 153 7 61 10 81 111 1 141 6 59 3 68 103 7 129 5 57 2 63 69 10 97 4 4 47 5 41 4 55 83 7 87 9 3 43 39 9 51 l 53 117 85 8 113 2 2 37 8 33 49 12 45 111 1 76 7 101 6 1 31 7 27 11 35 24 89 4 71 ll 99 5 115 3 73 12 1.500 2.000 2.500 3.000 3.500 4.000 4.500 5.000 5.500 1.999 2.499 2.999 3.499 3.999 4.499 4.999 5.499 5.999 Figure D12. Link Congestion Index Peak Half Hour Congestion Indices for Run H1. - Link Number Route Number LIST OF REFERENCES LIST OF REFERENCES Atherton, T. J., and Suhrbier, J. H., Urban Transportation Energy Conser- vation: Analytic Procedures for Establishing Changes in Travel, Final Report - volume II, 0.8. DOE, Office of Conservation and Ad- vanced Systems Policy, washington, D.C., DOE/PE/8628-1, 1979. Batty, Michael, "Recent Developments in Land-Use Modeling: A Review of British Research," Urban Studies, V01. 9, No. 2, June 1972, pp. 151-177. Betz, J. M., and Supersad, J. N., "Traffic and Staggered WOrking Hours," Traffic_guarterly, V61. 19, No. 2, April 1965. Bowman, L. A., Goetsch, D. H., and Polzin, S. E., A Model for Evaluating the Energy Conserving Potential of Transportation and Land Use Policies: Development and Preliminary Application, unpublished M.S. research report, Department of Civil Engineering, Northwestern Uni- versity, May 1975. Charles River Associates, A Disaggregate Behavioral Model of Urban Travel Demand, final report prepared for Federal Highway Administration, contract DOT-FH-11-7566, March 1972. Chang, M., Evans, L., Herman, R., and wasiekewski, P., "Gasoline Consump- tion in Urban Traffic," Transportation Research Record 599, 1976. Christopherson, P. D., and Olafson, G. 6., "Effects of Urban Traffic Control Strategies on Fuel Consumption," ITE Journal, November 1978. Claffey, P. J. and Associates, Running Costs of Motor vehicles as Affected bproad Design and Traffic, NCHRP Report 111, 1971. COMSIS Corporation, Traffic Assignment: Methods, Application, Products, U.S. Department of Transportation, Federal Highway Administration, washington, D. C., August 1973. Dupree, J. H., and Pratt, R. H., Low Cost Urban Transportation Alternatives: A Study of ways to Increase the Effectiveness of Existing Transporta- tion Facilities, Vblume I, Results of a Survey and Analysis of Twenty- one Low-Cost Techniques, 0.8. Department of Transportation, Office of the Secretary, January 1973. Edwards, J. L., Relationships Between Transportation Energy Consumption and Urban Spatial Structure, unpublished Ph.D. dissertation, Depart- ment of Civil Engineering, Northwestern University, June 1975. 201 202 Evans, L., Herman, K., and Lam, T., "Multivariate Analysis of Traffic Factors Related to Fuel Consumption in Urban Driving," Transporta- tion Science, Volume 10, No. 2, May 1976. Evans, L., and Herman, R., "Automobile Fuel Economy: An Alternative Interpretation of Recent Computer Simulation Calculations," Trans- portation Research, V61. 12, No. 2, 1978. 1 Evans, L., "How Does Traffic Speed Affect Urban Fuel Consumption," Letter to the Editor, ITE Journal, June 1979. General Motors Corporation, Truck and Coach Division, "vehicle Dynamics Simulation Model," Pontiac, Michigan, 1974. Goldner, W., "The Lowry Model Heritage," Journal of the American Insti- tute of Planners, volume 37, No. 2, March 1971. Greenberg, A. M., and wright, D. M., Staggered Hours Final Evaluation -- Queen's Park Demonstration, Ontario Ministry of Transportation and Communication, May 1975. Gross, J., Kocis, M., and Cohen, G., Preliminary Techniques for Estimat- ing the Net Energy Savings of TSM Actions, Planning Research Unit, New York State Department of Transportation, Albany, New York, October 1978. - Hartgen, D. T., "Energy Analysis for Urban Transportation Systems: A Preliminary Assessment," Transportation Research Record 599, 1976. Hatfield, F. J., "User Optimized Traffic Assignment," Transportation Engineering Journal, May 1974. Highway Capacity Manual, Highway Research Board Special Report 87, Nation- al Academy of Sciences, National Research Council, Publication 1328, washington, D. C., 1965. Holland, D. D., A Review of Reports Relating to the Effects of Fare and Service Changes in Metropolitan Public Transportation Systems, Center for Urban Programs, St. Louis University. Prepared for the Office of Highway Planning, Federal Highway Administration, U.S. Department of Transportation, Grant No. 32-22-66, June 1974. Honeywell Traffic Management Center, Fuel Consumption Study - Urban Traf- fic Control System (UTCS) Software Sgpport Project, Federal Highway Administration, Report No. FHWA-RD-76-81, February 1976. Jones, D. W. Jr., Nakamoto, T., and Cilliers, M. P., Flexible WOrk Hours: Implications for Travel Behavior and Transport Investment Poligy, UCB-ITS-RR-78-r, Institute of Transportation Studies, University of California, Berkeley, December 1977. LeBlanc, L. J., Mathematical Programming Algorithms for Large Scale Net- work Equilibrium and Network Design Problems, unpublished Ph.D. Dis- sertation, Transportation Center at Northwestern University, Evanston, Illinois, 1973. 203 le Clerq, F., "A Public Transport Assignment Method," Traffic Engineer- ing and Control, Vol. 14, No. 2, June 1972. Lieberman, E., and Rosenfield, N., Network Flow Simulation for Urban Traffic Control System -- Phase II, Volume 5, Extention of NETSIM Simulation Model to Incorporate vehicle Fuel Consumption and Emis- sions, Final Report, Prepared for the Federal Highway Administra- tion, Offices of Research and Development, Report No. FHWA-RD-77-4S, washington, D. C., April 1977. Lowry, I. S., A Model of Metropolis, memorandum RM-4035-RC, the RAND Cor- poration, Santa Monica, California, August 1964. Lutin, J. M., "Energy Savings for work Trips: Analysis of Alternative Commuting Patterns for New Jersey," Transportation Research Record 561, 1976. Masey, A. C., and Paullin, R. L., Transportation vehicle Energy Intensities, Ames Research Center, Moffett Field, California, NASA TMX-62,404, U.S. Department of Transportation, DOT-TST-13-74-1, June 1974. Mass Transportation Commission of the Commonwealth of Massachusetts, McKinsey & Co., Systems Analysis and Research Corp., and Joseph Na- politan & Assoc. Mass Transportation in Massachusetts. Sponsored by U.S. Housing and Home Finance Agency; Project No. MASS-MTD-l, July 1964. O'Malley, B. W., and Selinger, C. S., "Staggered werk Hours in Manhattan," Traffic Engineering and Control, Printerhall Limited, London. volume 14, No. 9, January 1973. . O'Malley, B. W., work Schedule Changes, Staggered WOrk Hours in New York, Port Authority of New York and New Jersey, August 1974. Owens, R. D., and Van WOrmer, G. H., "The Effect of Staggered WOrking Hours on Traffic at a Large Industrial Complex," Traffic Engineering, August 1973. Pratt, R. H., and Bevis, H. W., An Initial Chicago North Suburban Transit Improvement Program 1971-1975 - Volume II: Technical Supplement, Sponsored by the Urban Mass Transportation Administration, U.S. Department of Transportation; Project No. ILL-T9-2, July 1971. Pratt, R. H., Pedersen, N. J., and Mather, J. J., Traveler Response to Transportation System Changes: A Handbook for Transportation Planners, Contract No. DOT-FH-11-8479, U.S. Department of Trans- portation, Federal Highway Administration, Urban Mass Transportation Administration, February 1977. Peskin, R. L., and Schofer, J. L., The Impacts of Urban Transportation and Land Use Policies on Transportation Energy Consumption, U.S. Department of Transportation, Office of University Research, DOT- TST-77-85, April 1977. 204 Peskin, R. L., The Impacts of Urban Transportation and Land Use Policies on Transportation Energy Consumption, unpublished Ph.D. Disserta- tion, Department of Civil Engineering, Northwestern University, Evan- ston Illinois, 1977. Port Authority of New York and New Jersey, Staggered WOrk Hours Study -- Phase I Final Report, August, 1977. Quarmby, D. A., "Choice of Travel Mode for the Journey to work - Some Findings," Journal of Transport Economics and Policy, London School of Economics and Political Science, London, W.C.2, Volume 1, No. 3, September 1967. Remak, R., and Rosenbloom, S., Peak-Period Traffic Congestion: Options for Current Programs, NCHRP Report 169, 1976. Safavian, R., and Mclean, K. G., "variable WOrk Hours: Who Benefits?" Traffic Engineering, Institute of Transportation Engineers, Arling- ton, Virginia. Vblume 45, No. 3, March 1975. Sanders, D. B, Reynen, T. A., and Bhatt, B., Characteristics of Urban Transportation Systems: A Handbook for Transportation Planners, U. S. Department of Transportation, UMTA, UMTA-I IT-O6-OO49- 79-1, Whasington, D. C., June 1929. Santerre, G. L., An Investigation of the Feasibility of Improving Free- way Operation by Staggering Wbrk Hours, an unpublished thesis for the degree of M.S., Texas A & M University, Department of Civil Engineering, Houston, Texas, May 1966. Shunk, Gordon A., and Bouchard, Richard J., "An Application of Marginal Utility to Travel Mode Choice," Highway Research Record 322, 1971. Stopher, P. R., and Lavender, O. J., "Disaggregate, Behavioral Travel Demand Models: Empirical Tests of Three Hypotheses," Proceedings, Transportation Research Forum, Oxford, Indiana, pp. 321-336, 1972. Talvitie, A., "Comparison of Probabilistic Modal-Choice Models: Estima- tion Methods and System Inputs," Highway Research Board Record 392, 1972. Tannir, A. A., The Impacts of Feasible WOrk Hours and Compressed work- week Policies on Highway Networks, Transportation Economics, Orga- nizations and Employees, New York State Department of Transporta- tion, Preliminary Research Report 129, August 1977. Texas Department of Highways, Amarillo Urban Transportation Plan, Austin, Texas, 1964. United States Congress, Office of Technology Assessment, Energy, The Economy and Mass Transit, December 1975. 205 Voorhees, A. M., and Morris, R., "Estimating and Forecasting Travel for Baltimore by Use of a Mathematical Model," HighwayiResearch Board Bulletin 224, washington, D.C., 1959. Voorhees, A. M. and Associates, Inc., Factors and Trends in Trip Lengths, NCHRP Report 48, 1968. Voorhees, A. M. and Associates, Inc., Guidelines to Reduce Energy Consump: tion through Transportation Actions, U.S. Department of Transporta- tion, UMTA-IT-06-0092-74-2, May 1974. wardrop, J. G., "Some Theoretical Aspects of Road Traffic Research," Proc. Institute of Civil Engineers, Vol. 1, No. 2, Part II, June 1952. West, D. T., "A Technical Report of the 1974 Union 76 Fuel Economy Tests," Automotive Fuel Economy, Progress in Technology Series, vol. 15, Society of Automotive Engineers, 1976. Witkowski, J. M., and Taylor, W. C., "Urban Transportation Planning Under Energy Constraints," Transportation Research Board Record 707, 1979.