USF Libraries
USF Digital Collections

Decision support systems applications in urban transit systems management

MISSING IMAGE

Material Information

Title:
Decision support systems applications in urban transit systems management
Physical Description:
Book
Language:
English
Creator:
Nnaji, Soronadi
Mtenga, Primus
United States -- Dept. of Transportation. -- Research and Special Programs Administration
United States -- Dept. of Transportation. -- University Research Program
Publisher:
FAMU/FSU College of Engineering, Dept. of Civil Engineering
Place of Publication:
Tallahassee, Fla
Publication Date:

Subjects

Subjects / Keywords:
Local transit -- Management -- Data processing   ( lcsh )
Decision support systems   ( lcsh )

Notes

Statement of Responsibility:
principal investigator, Soronnadi Nnaji ; co-principal investigator, Primus Mtenga.
General Note:
"December 1994."
General Note:
Sponsoring agency: Office of Research and Special Programs, U.S. Dept. of Transportation.
General Note:
Supported by a grant from the U.S. Dept. of Transportation, University Research Institute Program.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 32933618
usfldc doi - C01-00076
usfldc handle - c1.76
System ID:
SFS0032194:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nmm 22 Ka 4500
controlfield tag 008 s1994 flu|||| ||||||eng
datafield ind1 7 ind2 024
subfield code a C01-00076
2 doi
035
(OCoLC) 32933618
040
FHM
1 100
Nnaji, Soronadi.
0 245
Decision support systems applications in urban transit systems management.
n Volume 1 /
c principal investigator, Soronnadi Nnaji ; co-principal investigator, Primus Mtenga.
260
Tallahassee, Fla. :
b FAMU/FSU College of Engineering, Dept. of Civil Engineering,
1994.
500
"December 1994."
Sponsoring agency: Office of Research and Special Programs, U.S. Dept. of Transportation.
Supported by a grant from the U.S. Dept. of Transportation, University Research Institute Program.
650
Local transit
x Management
Data processing.
Decision support systems.
700
Mtenga, Primus.
710
United States.
Dept. of Transportation.
Research and Special Programs Administration.
United States.
Dept. of Transportation.
University Research Program.
049
FHmm
8 773
t Center for Urban Transportation Research Publications [USF].
4 856
u http://digital.lib.usf.edu/?c1.76



PAGE 1

National .,QIL Urban .JOQ\. Transit -ar-ar-Institute ot t h e CENTE R FOR URBAN TRANSP ORT ATION R ESEARC H o f Sou t h Flotida Sta t e University Ftorida A & M Univers ity fk>rid a lnternat:onal University DECISION SUPPORT SYSTEMS APPLICATIONS IN URBAN TRANSIT SYSTEM S MANAGEMENT Volume 1 Principal Investigator: Soronnadl Nnajl CoPrlnclpattnvestlgator: Primus Mtenga Aorida A & M Univer s i ty December 1994 FAMU/FSU College of Engineering Department of Civil Engineering Tallahassee, F lorida, 32310 Phone: (904) 487 6127

PAGE 2

TRCIOOCAL REPO:RT STANO AltO 1TTLE PAGE 1, Report No. 3. NUTI93FAMU-2 ... .. DECISION SUPPORT SYSTEMS APPLICATIONS IN URBAN TRANSIT December 1994 SYSTEMS M ANAGEMENT (Volume I) 1. Al.lhottt) Dr. Soronnad i Nnap and Dr. Primus Mtenga 8 &. NemtWMftw 10. WOIICUf'llNo. Florida A & M U n lversl\y, FAMU-FSU College ol Engineering, 2525 Pottsdamer Street, Tallahassee, FL 32310 u DTRS 93-G-Q019 13. ACipOI! Office of Research and Special Programs July 1993 through Oc1ober 1994 U S. Departme nt of Transportation Washi n gton, D .C. 20690 t4. Supported by a grant from the U.S Department of Transportation, University Res earch lnstHute Program 1&. The Florida DOT list of transH objectives and measures of eHec11veness were combined wHh suggested objec1ives and measures from the IHerafure and case studies to develop a comprehensive set. A generic decison sHuatlon, def i ned to Include a goal, t hese set of objectives and sub objectives, and a set of was struc1ured in a hklrarchy wHh the goal at the top and the set ol aHematives a t the bottom. 'Expert Choice (EC) a decision support sys1em software packag e was adopted fo r analyzing and choosing between aHematlve courses of ac11on i n a transH decision sHuatlon. The generic decision hierarchy was coded in the software as a template. For a given decision sHuat ion an EC Model was generated from this templa t e by entering the goal and the particular to the sHuation. A broad spectrum of transH decision sHuatlons were identified from research IHerature and published TransH Development Plans (TOP) EC was applied to a seiec1ed number ol these decision sHuations The mode l input were the we i ghts calculated by EC using importance measure values assigned to the objec11ves at the same level of the hierarchy and t h e l i kelihood. preference or partonnance measure va l ues for assess ing how well the meet the lowest level objectNes. Nonj>&rtinenl objectives were assigned zero weights In J!le model. The values of the measures were abs1racted from the TOPs. The EC software as applied in this projec1, presents a very flexible and powertul tool for the analysis of transH decisio n sHuations and for prioritizing aHemallve courses of action developed to meet transH goals and objec t ives It contalns features for performing a variety of S
PAGE 3

TABLE OF CONTENT List of Figures .. .. .. ..... .................................................... .... ... .... .... . ... . . .... ..... .... .. ..... .... List of Tables .. .. . ... ... ... . ... ... . .. . .. . . .. ... ... . . . .. .. . . . . . . . . . .. .. ... . ... ... .. . . . . ... . ... .. . .. . i Acknowledgements . . . . .. .. .. . .. . . .. . . . ... .. . .. ....... .. .. .... ... .. ... . . ..... .. .. .. ... . .. .. . . .. ..... .. il Abstract .. ............ ... ............................... .......... .. ............ . ............... . .... ........... .. ........... .. iii Executive Summary 1 1. DECISION MAKING I N TRANSIT MANAGEMENT 1 1 Problem Statement . . . . .. .. . . .. . ... .. .. . ... . ... . ... . .. . .. .. .. . ... . .. . . . . . . . ..... ... . . . .. 3 1 .2 TranM management Activities .. .... . ..... . ... .......................................... .. ... ........... .... .. 3 1.2.1 The Decision Environment ... ... ... ........... .... .......... .......... ............................. ... 3 1 .2. 1 The Decision Situation Defined . ......... ............. ... ............ ........ ..... .......... ........ 4 1 .3 Transit Objectives and Measures of Effectiveness. .. ................................................. . 4 1 3. 1 Decision Objectives ........ .. ... .................... ............. ... .............. ..... ... . .... .... .. .. 4 1.3.2 Measures of Effectiveness ............... .... ... .......... .... ....... ... ......... ....... ........ .... .. s 1.3.3 Objectives and Meassures Used ...... ........ ... ... ....... ... .... .... .... ... .... .. .... .. ... .... 7 1 .3.4 Generic Transit Decision H ierarchy .. . . ............. .... ....... .................................. 7 1 3 5 Data Needs ... ... .... ...... .. .. ...... .. ......... .. ..... . ... ............................. .......... . ....... 7 2. DECISION SITUA TIONS 2.1 T ransit Development Plans Reviewed ......... ..... ....... .... ........ ..... .......................... . 12 2.2 Generic Case Studies of Decision Situations Identified ...... ......... ...... .... ....... ........... 13 3 DECISION SUPPORT SYSTEMS T!;C HNOLOGY 3.1 App lications In Transit management ............ ................. ............. ..... ..................... ... . 22 3.2 Decision Support Systems S oftware. .... ..... ... .... .. ..... .... ... ....... .......... ...... ........ .. .. ... 23 3 2.1 System Selection ................................ ... ... .. ... ..... .............. .... ... ......... ..... ..... 23 3 2.2 Development of the Generuic model .... ......... ........... ................................... .... 24 3.2.3 Using the generic Model 24 3.2.4 Assessments, Ev aluation and Analysis ... ... ...... .... ..... ........ ...................... .. .... 25 4. CASE STUDY APPLICAT IONS 4.1 Case Studies from Volusia Transit Development Plan ..... .. ............. ..... .............. ....... 32 4.2 Case Studies Synthesized from TPD's ... .. . . ... .... .................... ... ... ......... ... . .. .. . .... ..... 37 5 CONCLUS IONS .. . .. . .. .. .... . . . .. .. . .. . .. . . .. ... ........ .. .. 48 6. APPENDIX General In struction for C r eating an EC Mode l ...... ..... ... ......... ... ........ .... ... ......... .. ....... 51

PAGE 4

Figure 1.1 3.1 3.2 3.3 4.1 4.2 4.3 4.4 4.5 4.6 liST OF FIGURES Page Generic Transit Decision Hierarchy .. ...... ... . ............. ..... ... ....... .... .. ..... .... 9 Example of Pairwise Compaslson Using Numer i cal matrix....... ... . ..... . ..... 28 Example of Pairwise Compasison Using Graphical Comparison .... ..... ... 29 Example of Alternative Weights Obtained by Entering Hard data .... ..... ....... 31 Service Extension: Decision Hierarchy and Weights .................... .... .. ..... 38 Service Extension: Evaluat ion of Alternatives .... ........ .... .. ... .......... .... .... 41 Srvice Extension: Performance Sensitivity ... ........................ .. ... .... ... .... 42 Weekend Service: Decision Hierarchy and We i ghts ... .... . ..... ....... ........ ... 43 Weekend Serv i ce : Evaluation of Alternatives ......... ................ ... .... ... ..... ..46 Weekend Service: Performance Sensitivlty .. ... ............ ....... .. ....... ...... .... 47 LIST OF TABlES Table Page 1.1 Rorida DOT Transit Performance Measures........ .... .... ...... ......................... .. ... 6 1.2 Proposed Transit Objectives and Measureds of Effectiveness ...... .... ... . ..... .. ... 8

PAGE 5

ACKNOWlEDGMENTS This project was made poss ib le through a grant from the u. S. Department of Transportation, University Research I nstitute Program The support Is gratefully acknowledged. Mrs. Belinda Morris and Mrs. ldella Gallon worked on the project as Administrative Assistant and Fiscal Officer,. respec tively. The following undergraduate students worked on the project: Olivia Boykin, Scott Eddy, Harry Raysin and Nigel Richardson. The following graduate students worked on the project: Olukayode Adegoke, Janaki Hari and Chika Nwanna. MASTERS DEGREE PROJECT Chika Nwanna : Relative Performance of Hot In place Recycled Asphalt on Florida Roads" May 1995. ii

PAGE 6

ABSTRACT The Florida DOT list of transit objectives and measures of effectiveness were combined, with suggested objectives and measures from the literature and case studies, to develop a comprehensive set A generic decison silualion, deli ned to include a goa l these set ol objectives and sub-objectives, and a set of alternatives, was structured In a hierarchy with the goal at the top and the set of alternatives at the bottom. Expert ChOice (EC), a decision support system software package, was adopted for analyzing and choosing between aHemative courses of action in a transit deci sion s i tuation. The generic decision hierarchy was coded In the software as a Generic model. For a given decision situation, a Case model may be generated from this Generic model by entering the goal and the alternatives particular to the situation. A broad spectrum of transit decision situations were identified from research literature and published Transit Development Plans (TOP). EC was applied to a selected number of these deci sion situations. The model input were the weights calculated by EC using Importance measure values assigned to the objectives at the same level of the hierarchy, and the likelihood preference or performance measure values for assessing how well the alternatives meet the lowest level objectives. Non-perti nent objectives were assigned zero weights in the model. The values of the measures were abstracted from the TOPs. The EC software, as applied in this project, presents a very fleXible and powerful tool for the analysis of transit decision situations and for pr i oritizing alternative courses of action developed to meet transit goals and objectives. It contains features for performing a variety of sensitivity analyses. Persons involved In any facet of transit decision making will find the Generic model developed In the course of this project very useful. A runtime version of this model has also been developed and iS available for a nominal fee.

PAGE 7

EXECUTIVE SUMMARY The overall goal of a public transit system is to transport the citizenry to their destination in a safe, timely, comfortable and reliable manner, and to minimize pollution, congestion and energy consump tion i n the process. The cost-effectiveness of the system depends on the quality and quantity of the data used to make transit decisions. The goal of the project was to Identify the data needs of different transit decision situations and to examine the applicability of decision support system technOlogy to these situations. The individual components of the goal comprise, In general, the set of objectives. Some of these objectives Involve conflicting interests of various groups. The Transit Systems Performance Measures developed lor the Florida Department of Transportation by the University of South Florida's Center for Urban Transportation Research In cooperation with t he Florida Transit Association and Florida Transit Systems, were combined with suggested objectives and measures from the literature, to evolve a comprehensive set of performance measures. The data needed to execute the decision making process are the values of these measures. Most of the measures are non monetary ana methodS to quantify them aocurately and consistently in monetary terms often have serious limitations. Expert Choice (EC), a decision support system software package, was adopted for analyzing and choosing between alternative courses of action in a transit decision situation. EC Is based on the Analytical Hierarchical Process (AHP) -a powerful and comprehensive methodology that provides the analyst the ability to incorporate both qualitative and quantitative factors In the decision making process. Following the AHP methodology, a generic decison situation, defined to include a goal, a comprehensive set of transit objectives and sub-objectives, and a set of alternatives, was structured in a hierarchy with the goal at the top and the set of alternatives at the bottom. This generic hierarchy was coded in the software the GeneriC model. For a given decision situation, an EC Case model was generated from this generic model by entering the goal and the alternatives particular to the situation. The Information used by EC Is the weight estimate for the members at each level of the decision hierarchy For all but the lowest level objectives, the measure is considered to be the relative importance of the objective. For the lowest level objectives, the measure may be the likelihood of the alternatives meeting the objectives, the relative preference for the alternatives with respect to the objective, or it may be an attribute of the system that indicates the performance of the alternatives in meeting the objectives. All the Florida DOT measures are of this last category. The values for the importance, likelihood and preference measures are the user's judgements In the form of pair-wise comparison of the objectives. The judgement may also be numerica l values assigned to each objective. The judgements may be based on intuition, experience, or hard data. I

PAGE 8

Values for the performance measures are usually har
PAGE 9

1. DECISION MAKING IN TRANSIT MANAGEMENT 1.1 PROBLEM STATEMENT The overall goal of a public transit system is to t ransport the citizenry to their destination In a safe, t imely, comfo
PAGE 10

transn system and Incorporating elements of such plan into a t ransit development p lan (TOP). The aut hors point e d out that ther e is n o single co rrect formula for carrying out strategic planning but recommended a checklist of related processes and Issues prepared by TAB's committee on s trategic management as a starting point. The checklist include organization' s mission, environmental scanning, market analysiS, strengths and c r itical issues and strategies, strategic management The IDP framework discussed by Boyte and Ouderkirk is based on the methodology developed by the Florida DOT. Jensen -Flsher (1993) discussed the data needs of transit systems for both near and tong-term p lanning and deciSion making. The paper pointed out that many transit agencies collect all types data but lack the ability to analyze the data collected. There is, ther efore, a clear need for methods and techniques to transform tt1e volumes of data now available into formats that can serve effective route planning. 1.2.2 The Decision Situation Defined A decision situation i s defined to be a collection of a goal, a set of objectives and sub-objectives, and a set of alternatives. For the purposes of this study, each decision situation was described as a hiera rchy of task levels starting with the goal (Ieveii) through the alternatives (bottom level). A restriction on the hierarchic arrangement Is that any e lement i n one level m u st be capable of being related to some e lement In the next higher level, which serves as a criterion for assessing the relative impact of elements In tile level below. Decision making situations may be categorized according to the degree of structure. Structured decision situations Include those that do not invol ve a manager or otherwise follow standard operat ing procedures, semi-structured decisions situations include those in which manager ial judgement alone will not be adequate to sotve the problem; unstructured decisions include those that either are not able to be structured or that have not yet been examined In depth and so appear as unstructured. Transi t decision situations reviewed for thiS study fall primarily In the semistructured category. 1.3 TRANS I T OBJECTIVES AND MEASURES OF EFFECTIVENESS 1.3.1 Decision Objecti ves A comprehensive goal of public transit could be (Grigg, 1988, p 284) "to maintain a public transit system that provides access to places where citizens want to go in a safe, quick, comfortable, pleasant, convenient and reliabl e manner, and that helps minimize pollution, congestion and energy consumption i n the community." Goals for individual situations would, of course, be expected to be less comprehensive. Management objectives are the primary entities governing the directions in w hich the agency searches for tile best solution. They define what is or I s not to be accomplished In accordance to the 4

PAGE 11

specifications and constraints pert i nent to each decision s i tuation We have i dentified five broad transrt objectives to include: maximize service, maximiZe productivity, opti mize means and efficiency maximize patron acceptance, and maximize envlrornental soundness. Under these broad objectives are lower level objectives such as service quality, service consumption, capacity utilization lor serv i ce maximization. Further, lower level objectives may be identified depending on the management problem at hand. Typically, at any given level, the objectives will have different degrees of importance to the decision maker The alternative solutions under conside r ation are judged by the extent to which they meet the lowest level objectives Addressing t ransit decision situations from this multilevel approach has been lacking in practice as well as i n the literature. 1.3 2 Measures of effectiveness The Urban lns!Hute and International City Management Association (1974), (UIICMA) published measures of effectiveness lor basic municipal services including transit. For each municipal serviCe and objective, the report Identified quality characteristics, specific measures of the charactristics and suggested procedures lor collecting the data lor quantifying these measures. A summary of the principal measures is Included as Appendix A pages 284-286 in Grigg (1988). Zal
PAGE 12

Table 1.1 Flo rida DOT Transit Perfonnance Measures PERFORMANCE I N D ICATORS EFFECTNE MEASURES EFFICI ENCY MEASURES POPULA TI O N SERVICE SUPPLY FARE Cou n ty Po p ula tio n '"Vehiicle Miles per Capi t a Average Far e PASSENGERS AVAilABILITY ENERGY U T I L IZATION Passenge r Trips Re'J'90Ue Mi les PH R oute Mile Veh i c le M ile s per Gallon Passenger Milts O ays/Ho urs SetVice Availab l e VehiCle Miles per Hour VE H I C L ES SERVICE CONSU M P T ION OPERATING RATIOS Vehictes Available in Max. Service Pa ssenger Trips per Retmue Mile Farebox Recovery "Veh ic les Operated in Max. S9:f\lice Pauenger Trips per Revenue Hour l oca l Reven ue per Operating E xpense Spar e Ra tio "Passenger Trips per Capita "Opera ting per Operating Expense M i l ES/HO UR S OF S E RVI CE OUA UTY OF S E RVICE COS T EFFI C I ENCY V ehi d e Miles A\erage Speed operating Expense per Capita "Re venue Miles Average Age of Fleet Mai nt e n ance fxpeMe por O peraling Expen H V ehic r e H ours N umber o f Operaling Expense per Peak Vehicle Revenue Hours T o tal N umber of Roadcalls operating Expense pet Passenger Trip "RoU1e Mil es Rev enue Miles Between Accidenls Opera1 i ng E;xpens per Mile EXPENSES Revenue Miles Between Roadcalls Operating Expense pe1 Revenu e Mit "Total Opera!Jng Opeta1lng E xpense per Revenu Hour Total Maintenance EJq>enSe Maintenance Expens e per Revenue MISe Tot a l Capi ta l E"pens.e VEHICLE UTI L I ZAT ION REV E NUE S Vehi cle f ,iie$ per Peak Vehicle Total l ocal Reven ue Ve hicle Hcus per Peak Vehicle O p erati n g Revenue Revenue M!fts per Vehicle Mile$ Passenger Fare Re-.enue "Re'l'enue Miles pe r T ota l Ve hicles EMP L O YMENT Revenue Hours T ot.al Vehic l es Tot a l LABOR PROOUCTIVIlY 'rr3 nsp orta U o o Operating Employee Hours per Employee Maintenance Emoyee R e vtnue Hours per Employee A<%m! n ls tra ttve Employee Revenue H ours per M a lntenanoe Employee ENE RGY Revenue Hours per Admini$tfative Employee T o l al Ga ii OM consumed . Miles per Maintenance Employee Kilowatt Hour s o t Propulsion Power *P assenge r 'J"rips per Employte MeMures IO be reported in local TolaJ Vehicfes per Maint enance Emp1oyee 6

PAGE 13

1.3.3 Objectives and measures used The measures contained in Table 1.1 were combined with suggested objectives and measures from the UIICMA report to evolve a comprehensive set of performance measures. This set of suggested measures are given In Table 1.2. In this table, the upper case entries are the primary objectives, the bold faced lower case and the regular entries are the lower level objectives while the italized entries are the measures. The fourth column in Table 1 2 addresses system performance from the user vantage point while the fifth column provides an environmental consideration For all but the lowest level objectives the measure Is considered to be the relative importance of the objective. For the lowest level objectives, the measure may be the likelihood of the a lternatives meeting the objectives, the relative preference lor the alternatives with respect to the objective, or it may be an attribute of the system that indiCates the performance of the alternatives in rneetlng the objectives. 1 .3.4 Gen e r ic Transit Decision Hierarct)y The entries in T able 1.2 were translated into a hierarchy (Rgure 1.1) which we have labelled a Generic Transit Decision Hierarchy" since we expect that it Incorporates most of the objectives which one may be expected to encounter In the course of transit decision making. Abbreviations have been used, where necessary, to represent objectives and measures In the hierarchy. The key below the figure provides the definition f or each representation. 1 .3.5 Data Needs The data needed to execute the decis i o n making process are the values assumed by the above measures As may be noted from Tables t. t and 1.2, the measures have diverse units. For our purposes, the values for the importance, likelihood and preference measures are judgemental. The judgements may be abstracted from questlonaires and surveys or from interviews. They may a lso be numerical values assigned to each objective The judgements may be based on intuition, experience, or hard data. Va lues for the performance measures are usually hard data. They may atso be judgemental. An l mpcrtant source of data should b e the data collected and reported by public fransij providers in accordance with the FTAIFDOT repcrting requirements. For our purpose, the data requirements are those necessary to p rovide-values for the measures contained in Table 1 2. 7

PAGE 14

Table 1 2 Proposed Transit Objectives and Measures of Effectiveness MAX I MIZE SERVICE MAXIMIZE PRODUCTIVITY OPTIMIZE MEANS AND EFFICIENCY Service Qua lity Maxi mize Equip./Facility Utilization No. and frequency of accidents -Passenger trips/employee Vehicle hours per peak vehicle No and frequency of road calls -Revenue hoursladmin. employee Venic/e mites per peak vehicle Avera
PAGE 15

Fi 1.1 Generic Transit Decision Hierarch -E ACCOE N T S -QUAUTY RDCALLS AVSPEEO--f PATRONGECONSUME. FREQNCY CAPUT L N -MA.XSERV ...ISERTIME-SPANSERV COVERAGE SUPPLY-1?ASSPEM REVPAEMMA.X?ROO R EVPMEM r.'AXACCP REVPE.MLREVPOEM --{V HHPPVVHMPPV F C TYUT LN RVM PVM OPCOS T RVHPVMREVI\IH -OPCAPT -rCE?HR LOePML REVENUErOEPPT P ASSCST _. OEPPM r MEPRML MA/'IICST -.L M E POE rGASOP-ENERGY ---L ELEOP --FARE----E NUSERSiP COMPLS USERSATP -SEATNG-iN USERSTRCONRELS USESA.TR -RELBLTY MXPACC-RPMOVE USERSATT -ENUSERSIT-TRVTIMEMAXEVNt.l --r NOSPOl N LAtRPOLNj Abb re v iation Definiti o n ; I ACCDENTS I A IRP O L N J A LT-1 -ALT-2 A L T -3 Number and frequency o f aociden ls M i n i m i z e a i r pollution 1 Alte m at tv e No 1 Altemat i ve No. 2 Alt ern a tive No.3 j Alternative No. 4 9 \ ALT-1 ALT 2 ALT-3 AlT .. AlT-5 AlT-6 ALT-7 ALT-8 I ALT-9 j i I II

PAGE 16

Fig. 1.1 Generic Transit Decision Hierarchy (continued) ALT-44 No. 44 IALT-5 Altemabve No. 5 t; A. .O Al!em.s.t1Ve No. 6 --' ALT-7 AEte r native No 7 ,ALT..S AJiernative No. 8 : Alt ernative No. 9 : AV:;PEED Av erage Operating speed ICAPUTL N Capacity u til ization : Passanger t"'" per reenue mile ICOMPLS Maximize comfort an<1 pleasantness i C O NRELB Maximize convenence ar\d reliability CO NSUME I M a ximize service consumption : COVERAGE Maxim1ze service area covered (revenue mles per route m i les ) .ELEOP E f ectr City COOS"Umed ENERGY Minimize energy ulllizat1cn FARE I fare struet1S8 FCTYUTLN I E quipmeni/F acit UbiiZllt'0LVRY Optimize mean s and efficiency o f servtce delivery ,PASSCST Min. Cos t directly re l ated IO u ser 10

PAGE 17

Fig 11 Generic Transit Decision Hierarchy (continued) I PASSPEM Passanger trips per Employee PATRONGE Patrona-ge as meas ur ed by passa r .ger trips per cap ita QUA LI TY Ma x i mize service qua l i t y ROCA U S M u mber an d freq u ency of roadcalls : Mis btwn calls/No.of calls RELB LTY Max 1 mize re l iabi l ity ( ad h erance to sche d u l e ) I I REVENUE Maximize reve nue I ( EVPAEM Revenue hours per Administrative emp l oyees I REVPEM L R e v e nu e hours per total emp l oyees REVPMEM R e venue hou rs per mainte n ance em p loye e I 1REVPOEM Revenu e ho vrs per O p erat ing Employee I j REVTV H Reven u e miles pe r tota l v e h icles I I RPMQVE ) Maximize rap i d movement (Trave l t i mes) l I RVHPVM 1 Reven u e ho u rs pe r vehicl e m iles I I RVMPVM J Revenue m i les per vehic l e m i les I SEATNG 1 sea ling a"ait abillty I I I I SER TI ME 1 Maxim i ze serv i c e spa n I I lsPANSER V I Maxim1ze times serv i ce i s ava i lab l e ! SUPPLY 1 Max.imr ze Su p p l ied veh i c t e miles per capita I TRVTI M E Mi n i mize trave l time be t ween key poi n t s ---i U SERSATP 1 Use r sa i sfaetio n with comfort an a pleasantness USERSATT 1 Use r sat i s t sfaction i n relat ion to travel time I US ESATR Use r sat i s f actions with coo. &rel. I VHHPPV Veh icle hours per pea k vehide I V H MPPV Veh i c le miles per peak vehi d e I TRVTIME Min i mize trave l time between key poi nts US ERSATP U ser satsfact ion with comfo rt and p l easantness USERSA TT Use r satis!sfaclion i n re l a tion to trave l t i me USESATR User satisfact i o n s w11h con & rel. I VHHPPV Veh i d e hours pe r peak vehicle I I I -' I 11

PAGE 18

2. A SAMPLING OF TRANSIT DECISION SITUAnONS 2.1 TRANSIT DEVELOPMENT PLANS REVIEWED As a requirement or the Federal Transit Administration and the Florida Department of Transportation, an transn providers Wllo wish to apply for or WISh to continue receiving State Transit Block Grant funds are required to develop and periodically update a Transi t Development Plan (TOP). The plan should describe the organization, mission, goals and objectives of the provider as well as the providers strategies, policies and initiatives for meeting the goals and objectives. Fourteen(14) Transit Development Plans developed by public transit providers in the State of Rorida were reviewed lor this study as possible sources of real transit decision situations and insights into current transit management practices in Florida. The more comprehensive plans may meet the information needs for applying the DSS software package. The plans reviewed are listed below. This list is complemented by the various literature reviewed I n previous sections which contained examples of transit decision situations. 1. Bay County T ransit Development Plan, September 1992, prepared for Florida Department or Transportation, prepared by Bay County Council on Aging, Bay Coordinated Trans portation, Panama City Urbanized Area Metropolitan Planning Organization 2. Broward County T ransi t Development Program 1991-95 Update, November 1992, prepared by the Broward County Transportation Planning Division 3. Citrus County Coordinated Transportation Development, September 1993. 4. Clly or Key West Transit Development Plan, June 1993, prepared for Florida Department of Transportation District VI, by CUTR, College of Engineering, USF. 5. lakeland Area Mass Transit District (LAMTD) Transi t Development Plan Update, 1994, prepared by LAMTD and the Lakeland/Winter Haven UrbaniZed Areas MPO. 6. Lee County Transit Transportation Development Plan, July 1993, Developed by the LeemAN Staff 7 Metro-Dade T rans i t Development Program, June 1994, Prepared by Metro-Dade Transit Agency. 8. Palm Beach County Transit Development Plan Update, June 1994, prepared by the staff of the Office of the Metropolitan Planning Organization In conjunction with CoTRAN. 9. Pasco County Five Year Transit Development Plan, 1994-1995 Annual Update prepared lor the Pasco County MPO. 10. Pensacola/Escambia County Five-Year Transit Development Plan, July 1992. prepared lor the Pensacola Urban Area MPO, by CUm, College of Engineering, Unlv of South Florida 11. Sarasota County 1994 Transit Development Plan, May 1994, Prepared lor the sarasota County Area Transit (SCA l) prepared by the Sarasota County Transit Department 12

PAGE 19

12. Taltran Short Range Transit Development Plan, November 1991, prepared for Taltran prepared by BRW, Inc. 13. Tri County Transportation Development Plan, 1992. prepared for the Tri County MPO, prepared by Weslin Consulting Services in association with Ilium Associates 14. Votusla County T ransit Development Plan, June 1993, prepared for Volusia County MPO by BRW, Inc., Manuel Padron & Associates. 2.2 GENERIC CASE STUDIES OF DECISION SfTUATIONS The following decision situations were synthesized from the TOP's reviewed and case studies encountered In the literature. The objectives, and with these the measures of effectiVeness, have been couched in the format of the generic decision hierarchy described in section 1.3. 2.2.1 Decision Situation No. 1 Goal Determine Methods to satisfy the requirements of Federal F inal Rule 49 CFR Part 27, Nondiscrimination on the Basis of Handicap in Financial Assistance Program (issued May 1986) Objectives: -Service availability and utilization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) Alternatives: 1. Make existing bus service wheelchair-accessible 2. Provide special paralransit service in lieu of accessible regular service 3. Combination of the above two alternatives in a "mixed system" 2.2.2 Decision Situation No. 2 Goal: New and extension of existing routes so as to provide service to existing or proposed residential areas and..trafflc generators not presently served. Objectives: Service availability and utilization (MAXSERV) Productivity (MAXPROD) -Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives: 1. Combine route A and B to extend service to new area 2. Combine route A and C to extend service to new area 3. Combine route B and C to extend service to new area 4. Extend route A atone to provide service to new area 5. Extend route B alone to provide service to new area 13

PAGE 20

6. Extend route C alone to provide service to new area 7. Provide a brand new route to serve new area 2.2.3 Decision Situation No. ;3 Goal: Set headway Headway is defined as the time Interval, in minutes, between two successive departures May be set by pol icy or by demand Objectives : Service availability and utilization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives: I. xl minutes headway during peak hours and x2 minutes headway off peak hours In all routes 2. xt minutes headway during peak hours and x2 minutes headway off peak hours in selected routes, with x3 minutes head way in rema i ning routes 3. xt minutes headway throughout the day in all routes. Note: x1, x2 and x3 can be varied ( say 15,30, 40, 45, 60, 90 etc. depending on prevailing circumstances), therefore increasing the number of alternatives 2.2.4 Decision Situation No. 4 Goal: Set optimum span of service. This is the time of day in which the service is available. Does not have to be uniform over the service area. Objectives Service availability and (MAXSERV) Productivity (MAXPROD) Means and efficiency of service deliVery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives: 1. ALT-1: XI AM to X2 PM for all routes, through out the week 2. ALT-2: XI AM to X2 PM for all routes in weekdays, and X3 AM to X4 PM for all routes in weekends and holidays 3. A L T-1: XI AM to X2 PM for all for selected routes, and XS AM to X6 PM for remaining routes throughout the week 4. ALT-2: XI AM to X2 PM on weekdays, X3 AM to X4 PM on weekends for selected routes, X7 AM to XB PM week days for remaining routes Note: XHo XB can be varied ( depending on prevailing c ir cumstances), 14

PAGE 21

therefore increasing the number of alternatives 2.2.5 Dec;s;on Sjtuatjgn NQ. :i Goal: Determine how to phase-in improvements to a transit system. Objectives: 1) proposed traffic generator may be delayed or accelerated, thus affecting how proposed improvements to the mass transit system are to be implemented. 2) funding of transit Improvements may be delayed. Service availability and utilization (MAXSERV) Productivity (MAXPROD) -Means and efficiency of service delivery (OPDLVRY) -Patron acceptance (MAXACCP) Alternatives: 1. ALT-1: Extension of service to new a rea/ expansion of maintenance facilities/ headway improvements/construction major tra nsfer facility improvement of bus stop shelters in that order 2 AL T-2: Expansion of maintenance facilities/extension of service to new area/ headway improvements/construction major transfer facility 1 Improvement of bus stop shellers In that order 3. AL T-3: Construction major transfer facility/expansion of maintenance facilities/extension of service to new area/ headway Improvements! improvement of bus stop shellers In that order Note: Depending on the specific needs of the particular transit authority, more parameters can be added to those listed above. Many other combinations are possible, thus resulting to many other alternatives. 2.2.6 Decision Situation NQ, 6 Goal: Set capital Improvement requirements and p r iorities Objectives Service availability and utilization (MAXSERV) Productivity (MAXPROD) -Means and efficiency of service delivery (OPDLVRY) -Patron acceptance (MAXACCP) Alternatives: 1. ALT-1: Increase rolling stock 2. AL T-2: Improve maintenance facilities 3. ALT-3: Construction of additional transfer facilities 4. AL T -4: Establish Park and commute infrastructure 15

PAGE 22

2.2. 7 Situation No 7 Goal: Determine the best way to conduct Fleet Maintenance Objectives Service availability and utilization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) Alternatives: 1 AL T-1: Own and run maintenance workshOp for routine ma i ntenance and major repairs 2 AL T-2: Own maintenance shop and contract all maintenance operations 3 AL T -3: Own and run small facility for routine maintenance and contract all major repair jobs 4 AL T-4: Contract all maintenance activities (own no shop) 2.2.8 DecisiOn Situal!oo t:lo. 8 Goal: Determine the best fleet re-fueling scheme. Objectives : Serv ic e availability and utilizat ion (MAXSERV) Productivity (MAXPROD) Means and efficiency of se!Vice delivery (OPDLVRY). Especially the loss of revenue mi l es Patron acceptance (MAXACCP) Alternatives: 1. ALT 1: Centrally located filling depot maintained by the transit authority 2. AL T 2: Several, strategically located filling depots controlled by the transit authority 3. AL T -3: Contact with a private filing station retailer with strategically located filling stations 2 2 .9 Decision Situation No. 9 Goal: Determine strategiC structure of transfer locations Objective : Service availability and utilization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives: 1. ALT-1: One central station only (All rou t es have to go through this point, i.e originate or terminate at this point) 16

PAGE 23

2. ALT-2: One major central station with several satellite transfer locations 3. AL T-3: A network of small transfer stations (no one dominant station) 2.2.1 0 Decision Situation No. 10 Goat: Determine best methods of supplying the public with transit schedule Objectives: Service availability and utilization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives: t. ALT-t: Printed schedules stocked In each bus 2. ALT-2: Printed schedules stocked In selected bus stops 3. AL T 3 : Staffed telephone lines In operation during service hours 4. AL T-4: A combination of the above three alternatives 2.2.11 Decision SituatiOn No. 11 central Goal: Determine how park and ride facilities will impact the transit system and downtown traffic Objectives: Service availability and utilization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) 2.2.12 Decision Situation No. 12 Goal: Determine best ways of improving managerial structure Objectives Service availability and utilization (MAXSERV) Productivity (MAXPROD) -Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives: 1. ALTt: Autonomous division (finance, operations, maintenance and public relations), working In cooperation (horizontal hierarchy) 2. AL T-2: Top down organization for entire system 3. AL T-3: Maintenance division managed separately with remaining in a top down organization 17

PAGE 24

2.2. 13 Decision Sltualion No. 13 Goal: Determine how the transit system will be affected by the sub-contracting of selected services Objectives: Service availability and utilization (MAXSERV) Productiv ity (MAXPROD) -Means and efficiency of service delivery (OPDLVRY) -Patron acceptance (MAXACCP) Alternatives: 1. AL T-1: Sub-contract operations, fueling, fleet maintenance and custodial services 2 ALT: Sub-contract fueling, fleet maintenance and custodial services 3. ALT-3: Sub-contract fleet maintenance and custodial services 4 AL T -4: Sub-contract custodial services only 5. AL T 5: Nothing Is sub-contracted 2.2.14 Decision Situation No. 1 4 Goal: Determine ways of Increasing transit system ridership Objectives Service availability and utilization (MAXSERV) Productivity (MAXPROD) -Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives : 1. ALT: Increase parking fees downtown 2 ALT-2: Introduce special buses only lanes and/or streets during rush hours. 3. AL T-3: Conduct massive educational campaigns on the societal benefits that will result from using the transit system 2.2 .15 Decision Situation No. 15 Goal: Determine the optimum route separation Objectives: Service availability and utilization (MAXSERV) Productivity (MAXPROD) -Means and efficiency of service delivery (OPDLVRY) -Patron acceptance (MAXACCP) Alternatives: 1. AL T-1: X1 miles In high population density areas, X2 miles In medium density areas, an X3 miles elsewhere X1, X2 and X3 will be varied (say 0 :25, 0.5, 0. 75, and 1) to form alternatives 18

PAGE 25

2.2.16 Decision SilualjQO No. 16 Goal : Determine the optimum route round-trip cycle time A. This decision situation will impact most the following areas in the evaluation scheme: Alternatives: Service availability and uti lization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) -Patron acceptance (MAXACCP) The alternatives for this decision scenario are: 1. ALT-1: X minutes round trip X will be varied (say 15,30,45,60, 75,90) to form numerous alternatives that are feasible fo r particular locality 2.2.17 Decision Siluatism !llo, 1 z Goal: Determine headway's in two interdependent loops ObjectiVes: Service availability and utilization (MAXSERV) -ProductiVity (MAXPROD) -Means and efficiency of service delivery (OPDLVRY) -Patron acceptance (MAXACCP) Alternatives: 1. A L T-1: XI m in utes for loop A and X2 minutes for loop B X 1 and X2 will be varied (say 30, 45,60) to form numerous alternatives that are feasible lo r particular l ocality 2.2.18 Decision Situation No. 18 Goal: Determine how to Improve the public image of the transit system Objectives Service availability and utilization (MAXSERV) Productivity (MAXPROD) -Means and efficiency of service delivery (OPDLVRY) -Patron acceptance (MAXACCP) 2.2.19 Decision Siluallon No, 19 Goal: Establish optimum fleet stze 19

PAGE 26

Objectives: Service availability and uti lization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives: 1. ALT -1 : X busses per capital X will be varied (say 1/10000, 1/8000, 1/4000, 1/2000, 1/1000) to form numerous alternatives that are feasible for particular locality 2.2.20 Decision Situation No. 20 Goal : Establish optimum vehicle replacement age Objectives Service availability and utilization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives: 1. AL T -1: Repl ace a vehicle when maintenance cost exceeds X1 dollars per month, and the down time exceeds X2 hours per month XI and X2 will be varied depending on local conditions, thus yielding numerous alternatives 2.2.21 DeciSion Situation No. 21 Goal : Determine optimum transit system ownership and operation Objectives: Service availability and utilization (MAXSERV) Productivity (MAXPROD) Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) Alternatives: 1. ALT-1: Local government owned and operated 2. ALT-2: Local government owned and privately operated 3. ALT-3: Transit Authority owned and operated 4. AL T -4: Transit Authority owned and privately operated 5. AL T -5: Privately owned and operated 2.2.22 Decjsjoo SituaJjoo No, 22 Goal: Determine the best ways of coordinating activities of adjacent transit systems. This is critical for situations where we have urban centers close to each other. 20

PAGE 27

Objectives: Service availability and utilization (MAXSERV) Productivity (MAXPROO) Means and efficiency of service delivery (OPDLVRY) Patron acceptance (MAXACCP) 2.2.23 DeciSion Situation Np. 23 Goal: Determine the best way of budget allocation amongst the various modes of mass transt modes under the Transit Authority. Objectives Service availability and utilization (MAXSERV) Productivity (MAXPROO) Means and efficiency of service delivery (OPOLVRY) Patron acceptance (MAXACCP) Alternatives: 1. ALT -1: X1 percent for bus system, X2 percent for light rail system and X3 percent for special services, such as service to the physicaly challanged. X 1, X2 and X3 will be varied, resulting in numerous alternatives 21

PAGE 28

3. DECISION SUPPORT SYSTEM TECHNOLOGY The term "Decision Support Systems" (DSS) was coined fo r that segment of information systems management primarily Intended for use In supporting complex decisions. The concept was that oss applied in Situations where the decision maker was immersed in an environment of complex issues and interrelated factors at such a level tha t is it excee
PAGE 29

Guenthner and Sinha(1982) developed a model for analyzing the impact o f shOrt-term operating policy changes on selected performance measures. The user of the model enters data describing the operation characteristics of the transit system and the model computes the system performance. The data may be current or proposed operating characteristics. Maze and Dutta (1983) demonstrated the use of computer-based simulations as a planning tool which can be used to build a symbolic model of a system on the computer. Once constructed, the model can be used to experiment with the system without dis r upting tile operation of the real system. Maintenance managers can use this tool to simulate maintenance planning by being able to (1) determine parts inve n tory requirements by predicting failures with the model (2) experiment with preventive maintenance policies and (3) predict peakS, valleys and steady states of failure volumes enabling the maintenance manager to anticipate his labor requirements and his needs lor contracting out repair workS to private shops. Collura and McOwen (1984) noted that the use of microcomputers is becoming prevalent in many areas of transportation Management Information Systems (MIS) functions were grouped into sJx functional categories administration, planning, monitoring and evaluat ion, operations management, materials and equipment management, maintenance and, financial management. The exiting MISS were evaluated in terms of their capabilities, limilations, ease of use and relative costs. The authors observed that most of the MISS reviewed were not comprehensive In that they served one or more MIS functions but not all the major management information needs. The MISS that were comprehensive were very expensive and were designed for use in large transit systems. The major deficiency observed by the authors was the absence of an affordable, comprehensive MIS application for small fixed route trans i t systems (30 vehicles or less) The paper p rovides guidance for managers of small fixed-route, fixed-schedule services considering the purchase of a microcomputer and the necessary software for management i nformation purposes. 3.2 DECISION SUPPORT SYSTEM SOFTWARE (EXPERT CHOICE) 3.2.1 System Selectioo Two primary types of systems may be identified -DSS's designed to help reach difficult practical deCisions and those intended pQnariliy for managing and reporting information. Of Interest in this study are the former type. For these, Davis(1993, p 13) listed the following necessary properties and characteristics: 1) The structure and environment of the problem is not rigid or constant but instead is susceptible to change. 2) The exchange (interaction) between the user and the computer Is lnteactive in nature and in a dialogue that is informative, tolerant, and does not cause apprehension to the user. 23

PAGE 30

3) The system provides tne user IM!h tne capability to examine different situations evaluate various scenarios, and answer a variety of whatif quesijons. 4) The user Is afforded the flexibtlty to adapt tne system to his preferences, or changes In condition, that influence the decision (changing resource levels, costing factors, Interest rates, terminolgy, policies, ele.) Several software packages are availbie in tne market place. The Expert Choice (EC) software developed by Expert Chobl Inc. was considered to meet the abOVe requirements and thelefore was selected for thiS project. EC IS based on tne AnalyticaJ Hierarchical Process (AHP) a powerful and comprehensive methodology that provides groups and individuals tne abtllty to Incorporate both qualilallve and QUantitatiVe factors in the decision making process ( Saaty1992), The AHP uses a hierarchical model comprised of a goal, objectives, perhaps several levels of sub-objectiVes and alternatiVes for each problem or decision. It Is a general method for structuring lntrlcale or HI-defined problems and is buHt around the principles of constructing hierarchies, of eslablishklg p!IOiities, and of logical consiStency. EC can accommodate a variety ol data types and merge them Into a single overall measure to detennlne which alternatiVe is lhe most des irable. It does this by devising a scale that enables the user to measure intangible qualities. 3 2.2 Development or me Geoerlc Model On Invoking ECWIN (Expert Choice for Windows) the user Is presented with a menu of Icons. Of Interest in model building and utilization are the Structuring and the Evaluation and Choice icons. A new model may be built starting from either icon. The general instructions for building starting from the Structuring and the Evaluation and Choice Icons are outlined in Appendix I. The arrangement of the generic declson Situation in a hierarchical fonnat was described In Section 1.3. An EC generic decision model was developed from U!e hierarchy following U!e steps in option A. Appendix 1 The Choice of option A was based on our experience with boltl options The above Generic model has been naiTII!{j TRANSIT.STR. 111e extension "STR" Is assigned automatically and Indicates that the model is in its structural form and needs to be converted into the executable form TRANSIT.EC1. In this form, referred to In the manual as the Evaluation and Choice model, H can then be used to apply judgements to the objectives and assess the alternatiVes. A runtime version of the Generic model has also been developed and is avail able for a nominal fee. 3.2.3 USing the Generic Model The user wiN need to edU this generic model to fit the particular decision situation of interest by c hanging the goal, the set of alternatives and, if necessary. the weights assigned to the objectives. We refer to the resulting model as the case model. It Is noted that the weights 24

PAGE 31

assigned to certain objective peers may be fixed as a policy matter. Noni)ertinent objectives should simply be assigned zero weights Th is will elfectively eliminate this objective and its descendants from influencing the ranking of the alternatives. Procedure a should be followed to create a new case model and procedlJ"e 'b" should be followed to edn an existing case model. a) Io <;(l!aJe a New Case Model. step 0: Copy the TRANSIT.STR file into the ECMODELS directory that comes with the ECWIN software package. Step 1: Start the ECWIN software and use the Structuring command found under the File menu to open the TRANSIT.STR file. Step 2 : Specify the goal of the decision sruation at hand, using the Goal specification command (under the Edit menu). Step 3: Click on the A button on the Icon menu to bring up the generic alternatives. Edij the aHematives to reflect the decision situation at hand. step 4: Save the edijed TRANSIT.STR using the Save As command (under the file menu) using a name other than TRANSIT.STR but retaining the STR extension. Step 5 : An EC Case model can now be built by invoking eijher the Build EC Model command Of the Evaluation/Choice command under the File menu. When the Build EC Model COI'nmand is used the model is created without leaving the struc1uring screen. The user wil need to exij the structuring model and restart ECWIN in order to acoess the newly created model. When the Evaluation/choice is used the newly created Case model will automatically appear on the screen. In both cases, the name of the created EC model will be the same as the name specified in step 4, with the extension .EC1. step 6. Assessments, analysis, evaluation and generation of reports can now be implemenled b) To edn an wQsting Case model. Step 1: Start the ECWIN software. Depending on how the software was tenninated on last use, ECWIN will eijher open to a blank saeen or to the previous EC model. If this model is the one to be edijed, continue to the next step, otherwise access the desired model using the Open command under the File menu. Step 2: Click on the goal node. Use the Node Name command under the Edit menu to edit the goal Step 3: Click on the alternatives, one at a time. Use the Node Name command to edit both the name and definijion of the alternative. After making all the changes, the user will be asked if the changes are to be global. The typical answer should be yes. Step 4: The modified model is now saved as a new file using the Save As command. Step 5: Assessments analysis, evaluation and generation of reports can now be implemented. 3.2.4 Assessment. Evaluation and Analysis Assessment includes the data and judgemenl enby activities in addit ion to the quantification of the verbal judgemenls entered. Evaluation involves the prioritization of the alternatives while 25

PAGE 32

Analysis involves analysis in various modes. a) To enter Judgements and Data The Pairwise Data and the Vllhatif commands are invoked fonn the Assessment menu to enter data into the model. Pairwise: This command is used to make pairwise comparison of the alternatives. There are three of comparison: Importance: Alternative X is more important than Alternative Y. Preference: Alternative X is preferable to Attemative Y Likelihood: Alternative X is more likely than aHemative Y and three modes of comparison: VE!fbal: Graphical: Numerical: Use nonnal, daily language adjectives to describe the difference Use pie charts and bar charts to compare aHematives pairwise. Using numbers i n a questionnaire or matrix format Data: This command is used to enter data associated with each aHemative. This data may be survey data of resuH of separate analysls. Whatif: This command a l lows the user to experiment with different numerical values for certai n altemative(s) while observing the behavior of the remaining a l temative(s) The infonnation used by E C Is t11e estimate for the members at each level of the decision hierarchy. For all but the lowest level objectives, the measure Is considered to be the relative irT4J(lrtance of the objective. For the lowest l evel objectives, the measure may be the likelihood of the alternatiVes meeting the objectives, the relative preference for the alternatives with r espect to the objectiVe, or It may be an attribute of the system that indicates the performance of the al t ernatives i n meeting the objectives. The valueS for the irT4J(lrtance, l ikelihood and preference measures are the users judgements in the form of pair wise comparison of the objectives. The judgement may also be numerical values assigned to each objective. The judgemenl& may be based on intuition, experience, or hard data Values for the performance measures are usua l ly hard data which may be entered directly into EC or imported from an external source such as a spreadsheet. In a l l cases EC calculates the weights from the values entered. "The main steps for entering judgments in the verbal compari son mode include: (1) Enter the Vert>al Comparison mode from the parent node of the group of nodes to becompared. (2) Enter the judgments for each pair I n this group. (3) .Calculate the priorities ancllnconsistency for the group of nodes that were compared. (4) Examine Inconsistenci es if necessary (when Inconsistency ratio Is more than 0.1 O), and improve consistency from within the NumeriC Matrix roode. (5) Return to the main screen and observe the priorities that now appear for the nodes you compared. (EC Help Manual)" 26

PAGE 33

Given below are examples of the use of the Pairwise comparison and Data commands. Fig. 3.1. presents a relative preference in a Pairwise comparison, obtained in a numerical matrix mode, of the alternatives in relation to user satisfaction with travel times in a transit system Fig. 3.2 presents a relative preference in a Pairwise comparison, obtained graphically, of the alternatives in relation to user satisfaction to convenience and reliability of the transit system. Fig. 3.3 presents weights of the alternatives In relation to user satisfaction with comfort and pleasantness of the system as obtained by data entry, i.e., parentage of users responding favorably. b) Evaluation a!N lhls command derives priorities (weights) from simple pairwise comparison judgments. It then synthesizes or combines these priorities to obtain overall priorities for the attematives at the bottom of the tree. This result not only Shows the ranking of the alternatiVes, but also provides a meaningful (ratio scale) measure of the differences between the alternatiVes. c) Pelformance..petJsiWJrt "Pelformance Sensitivity iS the default sensitivity analysis mode. When the user opens a model after entering Sensitivity from the Senslllvlty Icon In the Windows Program Manager, the Case model is opened in the Performance analysis mode. To enter Performance from inside Evaluation and Choice select Sensitivity-Graphs then Performance. The Performance graph puts all the information about how alternatiVes behave vis-ll-vis each objective (criterion) on a single screen. This is the most compact presentation of the informatiOn about the priorities in the case model. Performance provides a composite sensitivity presentation ShO'Mng hOw well each alternative performs on each crtterion and overall, when an tile criteria are taken into account. Each objective (criterion) is shown by a vertical line. The point where an alternative line intersects such a vertical line as read from the axis on the right (labeled "Ait"lo"), indicates the priority the alternative received on that objectiVe. The overall priority of each alternatiVe is where it intersects the rightmost axis. The priority of each criterion is Shown by the small blue rectangular box on t hat criterion's vertical ine, as read from the axis at the left Oabeled Crito/o). What-if analysis Is done by dragging the box up and down the vertical fine to change the priority of the criterion. As the priority changes, the overall priorities of the alternatives on the axis at the right change and the other priorities are readjusted to accommodate the changes. The lines for the alternatiVes between the vertical criterion lines have no meaning. (EC Help Manual)" 27

PAGE 34

Node: 43200 Compa r e the re l ative PREFERENCE w i th respect to : USERSATI < RPMOVE < MAXACCP < GOAL I i\\.T 2 ,\,U,:J A\. f 4 ALT.S AI.T 6 .\LT.J AI,.T .a A \1 ALT .1 2 0 ,. o o 2.0 I>L f I 2 0 m o o 20 < > 0 ALr.3 I I 20 20 > O ,, "" . I "'-o 10 2 0 AC 20 o .. ... o .. AlfT I I "'-'4 I .. .. D eliniUon A L T-1 Alternative No. 1 ALT-2 AHernative No. 2 ALT3 Al t e r nat i ve No. 3 A L T-4 Al t erna t ive No 4 ALT -5 A l ternative No. 5 A LT6 Alternative No. 6 ALT -7 Alternative N o 7 A L T8 ( Alterna ltve No. 8 ALT -9 J A1terna11ve No. 9 A L T 1 .257 ALT2 .180 ALT-3 136 A L T-4 107 A L TS .C-94 ALT-6 .068 ALT .066 ALT8 045 ALT9 047 I ncons i stency Ratt o =0 1 1 Fig 3.1 E x ample of Pairwise Comparison Using Numerical Matrix 28

PAGE 35

Nod e :42200 Compare the relative PRE F ER ENCE with respeclto: USESATR < CONRELB < MAXACCP < GOAL Fig. 3.2 Example of Pairwise compa riso n U sin g Graphica l Compari s on 29

PAGE 36

Abbreviation Definition ALT-1 Alternative No. 1 ALT2 Alternative No. 2 ALT Alternative No.3 AL T-4 Alterna t i ve No. 4 ALT-5 Al tern ative No 5 ALT-6 Alternative No. 6 ALT-7 Alterna t ive No. 7 ALT-8 A lt ernative No 8 A L T-9 Alternative No. 9 ALT 1 0 4 ALT .107 ALT 3 .102 ALT-4 .099 ALT-5 .089 ALT-6 113 ALT-7 .122 ALT-8 .111 ALT-9 .152 Inconsistency Ratio =0.05 Fig. 3.2 Example of Pairwise comparison Using Graphical Comparison (Continued) 30

PAGE 37

Node : 41200 Dat a w!th respect to : USERSATP < COMPLS < MAXACCP < GOAL ALTI 60 ALT 65. ALT-3 55. ALT-4 45. ALT-5 70. ALT-6 75 ALl/ 45. ALT 8 58 ALT-9 80 AbbrtvlaUon Definition ALT1 Alternative No. 1 ALT-2 Alternative No. 2 AL T -3 Alternative No 3 ALT -4 Alternallve No. 4 ALT-5 Alternative No 5 ALT-6 Alternative No 6 ALT-7 Alternallve No 7 ALT-8 Alternative No. 8 ALT 9 Alternativ e No 9 ALT-1 .108 ALT-2 .118 ALT-3 099 ALT-4 .081 ALT-5 .12 7 ALT-6 136 AL T 7 081 AL T -8 ,105 ALT-9 .145 Inconsistency Ratio =0. 0 Fig. 3.3 Example of Alternative Weights Obtained by Entering Hard Data 31

PAGE 38

4. CASE STUDY APPLICATIONS 4.1 CASE STuOIES AeSTRACTEO FROM THE VOI.USUI A COui'ITY TOP The County developed several transij service decision slluations based on identified transit needs olllle COITlflUlity. Several anematives were developed for each decision sijua tion. The reader should note 111a1 what we refer to here as decision sHuatlons were called alternatives i n the TOP. Also mat we refer to here as alternatives were called options in the plan. Thus Decision Situation 1 corresponds to AlternatiVe 1 in the plan. Version 8 of the Expert Choice DSS software was appfied to Decision Situations 1, 6 and 9. SijuatiOn 1 Objectives: response Alternatives: Data: Analysis: Transit Service administration and Ooeratjons Reassess the system administration and operational structure in to the fragmented nature of transit service provisions in Volusia County. 1 A: Local Government Owned and Operated 1 B: Local government Owned and Privately Operated 1 C: Transn AUthority Owned and Operated 1 D: Transit Authority Owned and Prively Operated 1 E: Privately Owned and Operated There is no data for ltlis alternative. The seven performance measures are the overal l goalS for the transn authority and are as follows: 1. Coordination Maximize coordination of transit services 2. ADA Ensure compliance with the ADA 3. Expansi on Establish aJramework for future expansion 4. Mobility Effrciently and effectively serve tile mobility needs of ltle communi ty 5. SeMces Provide transportation services to those without access to regular autos 6. Funding -COordinate and maximize funding for transit system 7. Forum-Address transportation and t ranslt issues Note: The list of overall goals is 8 with the omitted goal being to establish recommendations that have the potential to Improve cost effectiVeness of transit services. Due to ltle limitation of the decision software to 7 leaves (performance measure) ltlls goal was omitted. 32

PAGE 39

priority Results: Verbal Judg ements of I mportance for each measure a r e as follows: Coordination is moder ately more important tllat ADA, Mobility Services and Funding, and Is equal to Expansion and Forum. ADA Is moderately less il1'4l0rtant than Expansion and Forum and Is equal to Mobillly, Services, and Funding. Expansion is moderately more important than Mobility, Services and Funding, and equal to forum. SefVice Is equal to Funding and moderately less important than Forum. Funding is moderately less important than Forum. The follOwing describes the comparison for each alternative With respect to the local node and the overall goal: Coordination, Expansion, and Forum: 1C-Ov.l'led and Operated is most likely to meet lhe above measures and has a praedia of 0.097. 1 0 -Transn Authority Owned and Privately Operated has a of 0.061 1A. 18, and 1E has priorities of 0 .024. ADA, Mobility, Services, Funding: All allernatives are equal in i mportance With respect to meeting the bove measures and have equal priorities of O.ot5 The priorities for each measure are as follows: Coordination 0.231 ADA 0.077 Expansion Mobnity Services Funding Forum0.231 0.077 0 .077 0.077 0.231 1C is most likely to meet measures 1 ,3 7 and is equal to aU other alternatives for the remainder of measures. 33

PAGE 40

Conclusions: SensitMty : 1 D is nex1 most likely to meet measures 1, 3, 7 and is equal to all other for the remainder of measures. 1 A, 1 B, 1 E are equal In all respects to meeting the measures. Alternative 1 C is the best alternative to satisfy the estab!ished performance measures. All alternative changes are for increase in priority. Coordination ADA 1C Increases with increase in priority 1 D increases at lower rate than 1 C 1 A, 1 B, 1 E decreases 1 C decreases 1 D decreases at lower rate than 1 D 1A, 18, 1E increases Expansion-1 C increases with increase in priority 1D increases at tower rate than 1 C 1 A, 1 8, 1 E, decreases Mobility1 C decreases 1 D decreases at lower rate than 1 D 1 A, 1 8, 1 E. increases Services1 C decreases 1 o decreases at lower rate than 1 D 1 A, 1 8, 1 E increases Funding-Forum1 C decreases 1 0 decreases at lower rate than 1 D 1A, 18, 1E Increases 1 C Increases wltll increase i n priority 1 D increases at lower rate than 1 C 1A, 18, 1E decreases 34

PAGE 41

Situation 6: Extension of Service Along u.s. 1 North Options: Date: Analysis: Provide service connections to the U.S. Route 1 area north of Ormond Beach. Extend service to the industriaVhotel uses In the area near the municipal airport and the 1-95 interchange. 60 Minute Headway 30 Minute Headway The follOwing data Is provided for this alternatiVe: Vehicles Operating Cost Ridership Potential 60 Minute HeaclwS 1 $106,000 35,100 30 Minute Headway 2 $212,000 70,200 The t hree sets of numeriCal data were analyzed using EC-8 with the comparison ol date mode using the above listed data. Each of the three measurements had an equal priorities of 0.333. Results: The synthesis of the above data results In a J)f'lorlty value for the 60 minute headway option 0.556 and 0.444 for the 30 minute headway. Conclusion: Using the above data for comparison, the 60 minute headway is determined to be the best selection. Sensitivity: The 60 minute headway priority Increases while the 30 minute option decreases for both vehicles and operating cost. For the ridership priority, the 30 minute option increases and the 60 minute option decreases. A cross over point is located at approximately 50% for both options. 35

PAGE 42

Situation 9: AHematlve: Options: Data; Analysis: Results: Conclusions: Sensitivity; line Haul Circulator Combination for New Smyrna Beach and Establish a New Smyrna Beach circulator and provide coverage to same areas now covered, with possibility ol expansion to adjacent neighborhood areas. 1 Vehicle Circulator 2 Vehicle Circulator The lollowing data iS provided for thiS alternative: Vehicles Operating Cost Ridership Potential 1 vehiCle Circulator 3 $248,700 94,950 2 VehiCle Circulator 4 $322,600 106,490 The three sets of numerical data were analyzed using EC-8 with the comparison of data mode using the abOve listed data. Each ollhe three measurements had an equal priorities of o.333. The synthesis of the above data results in a priority value lor the 1 VehiC!e Cirou!atQr option of 0.536 and 0.464 for the 2 Vehicle Circulator. Using the above data lor cornperison the 1 Vehicle Circulator is determined to be the best selection. The option to be included In the recommended Hst of anernallves is the 2 Vetlk
PAGE 43

4.2 CASE STUDIES SYNTHESIZED FROM SEVERAL TOPS Version 9 of tile Expert Choice DSS software was applied to Decision Situations two decision situations below. Situation 1: Extend service to a new area. AlternatiVes: 1. AL T-1: Combine route A and B to extend service to new area 2. ALT-2: Combine route A and C to extend service to new area 3. ALT-3: Combine route Band C to extend service to new area 4. AL T-4: Extend route A alone to provide service to new area 5. AL T -5: Extend route B alone to provide service to new area 6. AL T-6: Extend route C alone to provide service to new area lhe assigned and derived weights of the pertinent primaJy and tower level objectiVes are inclUded In FIQUre 4.1. The evaluation of the aijernatives Is Included In Figure 42 while the Perfonnance sensitivity analysis is shown in F igure 4.3. Situation 2 : Detennine the optimum span of service on weekends. Assumption main traffic generators are shopping malls and worship services on Saturday evenings and Sunday mornings. Aijematives: 1. Morning to Worship Centers, Afternoon to Shopping/Recreation Centers, service span 6 a .. m. to 6 p .m. 2.Moming to Worship Afternoon to Shopping/Recreation Centers, Service span 8 a.m. to 10 p .m. 3. Unifonn service in all weekend routes starting at 6 a.m. and end at 6 p m 4. Unifonn service in all weekend routes starting at8 a.m. and end at 10 p .m. lhe assigned and derived weights of the pertinent prlmaJY and lower level objectiVes are included in FIQUre 4.4. lhe evaluation of the alternatiVes is included In 4.5 while tile Perfonnance sensitivity analysis is shown in Figure 4.6. 37

PAGE 44

ACCOENT$-(.029) QUAL I TY RDCALLs( ll67) ( .022) AVSPEED--(.006) PATRONGE-(.017) COMSU FREONCY-(.o62) (.021) MAXSE CAPlJTLN-( .2U) (.023) JSERTIME--( .028) sPANSE R COVERAGE;-( .047) (.019) SUPPLY-( .048) PASSPE._ ( 03) REVPAE'--( 002) MAXPROD-REVPM EM-( .182) ( .041 ) REVPEML-( 038) . REVPOH-(042) (.003) VHMPPII-( 009) FCTYUTLN-RVMPVM -( .05) ( .011) (.012) Fig. 4.1 Service Extension: Decision Hierarchy and Weights 38

PAGE 45

REVTVfol...-( 0 14) OPCAPTlOEPMLI RMLO N E ( 0 1 5) ( .008) ( 0 155) O EPHR-RBALONE (.008) (0. 169) REV ENUE--RCALO N E ( .021) (0.170) p O A L OPOLVRY-OPCOSTJOEPPT -CMBRA8 ( 173) ( .052) ( .005) (0 158) PASSCSl O E PPM -CMBRAC ( 01) ( .005) (0.188) J MEPRML-I CMBRBC (.008 ) (0.17 1) MANCST MEPOE( 015) ( .008) JGASOP-( 037) ENERGY E LEOP ( 037) ( ) FARE ( .025) NUSERSTP-COMPL USERSATP-( 0 5) ( 027) SEATNG--(008) NUSERSTP-( 0 1 2 ) MAXACC CONRE L USESATR-( . 233) ( 1 ) RELBLTYM X PACo--( .033 ) NUSERSTI-( 01 4 ) Fig 4 1 Service E x ten sio n : Decision Hierarc h y and W eigh ts (continu e d) 39

PAGE 46

RPMO"' Mode: USERSATT-l ( 083) l ( 032) I TRVTIME' (.037) I MAXEVN1 NOSPOL,_ I (.19 7 ) ( .126) ' I AIRPOlN-' <.On) I I ' I Abbrevia tion Oeflntuon ACCDENTS Number a n d frequency of accidents I AIRPOLN Minimize air pollution AVSPEED Average Operating speed I CAP U TLN Capacity utilization:Passanger trips per revenue m il e i CMBRAB Comb ine route A and B t o extend service to new area I I I CMBRAC Combine route A and c to extend service to new area I CMBRBC Comb ine route B and C to exte n d serv ice to new a re a COMPLS Maximize comfort a n d pleasan t n ess I CONRELB Maximize convenience and reliabilily I CONSUME Maximize service consumption COVERAGE Maximize service area covered ( r evenue m iles per route mil es ELEOP E l ectricrty consumed I ENERGY Minimize energy utilization FARE Optimize fare structure I FCTYUT L N EquipmenVFaci lity Utili zation FREQNCY Maxim i ze passenger trips per reve nue hour I GASOP Gas consuption I I MANCST Minimize Maintenace cost MAXACCP Maxim i ze Patron Acceptance MAXEVNM Maxim i ze environmen t al soundness -MAXPROD Maximize Productivity I MAXSERV Maximize Service i MEPOE Maintenance expense s per ope r ating expen s es I MEPRML Ma i ntenance expenses per revenue mile I MXPACC Maximize patron conveniece l NOSPOLN Minimize no ise pollut i on I Fig. 4.1 Service Extension: Decision Hierarchy and Weights (continued) 40

PAGE 47

CMBRBC .171 RCALONE .170 RBALONE .169 CMBRAB .168 CMBRAC .168 RAALONE 1 55 Abbreviation CMBRBC RCALONE RBALONE CMBRAB CMBRAC RAALONE Distributive Mode Definition Combine route B and C to extend service to new area Extend route C alone to provide service t o new area Ex tend route B alone to provide service to new area Combine route A and B to extend service to new area Combine route A and C to extend service to new area Extend route A alone to provide service to new area ____ ___ .. -------Fig: 4 2 Service Extension: Evaluationof Alternatives 41

PAGE 48

.SO .70 . . ''. . . / . . .30 Ch!BRBC RCALONE .80 r---,\, \ . // . .20 ABALONE .liO \ ..... / \ '. .... ... . -. .\... ... ,.. .. -:.: ::= :...: r < .... .. ...... _.,_,..,_ ...; r o:.. ____ ... ; .--. -.... .50 -\' . _;'/ 4 0 .30 .\ / .20 .10 .10 .OOILL--..i:L.:-:--....L.I...--...Ll::-:::---U----:::-::-L---:-___!.00 MAXPAOD MAXACCP OVERALL MAXSEAV OPOLVRY MAXEVHM Abbr e v iation Definition MAXSERV MAX PROD Maximize PlcductMiy OPOL VRY ; uptimu::t mean a end of service delivery MAXACCP Pal ron Aecqptance MAXEVNM Malllmlzt environmental soundness CMBRBC Combine roult Band C to el'lend service to n ew a r ea RCALONE E:dtnd routt C a ton e to provide service to new area RBALONE Edend route B atone to provide service to new area CMBRAB Combine routt A and 8 to extend service to new area CMBRAC Combine route A and c to extend ser"fice to new aru RAALONE txtend rou1t A atone 10 proYide servico to now area -Fig. 4 3 Service Extension: Performance Sensitivity 42 CloiBRAB OIBIIAC RAAlOIIE

PAGE 49

I OisltDu:; e Mode I I ACCDE NTS I (.022) I QUALITYRDCALL S I ( 0 48) ( 0 1 7) I AVSPEEO (.009) I PATRONGE -(.011) CONSU M E -FREONCY -( 046) ( 0 1 5) MAX.SERV CAP U TLN-(.16) ( 019) JSERTIME-( 0 2<) SPANSERV CO VER AGE (.041) ( 0 17) SUPPLY -( .025 ) PASSPE M ( .052) REVPAEM-MAXPR O D -R VPMEM ( .24) ( .047) REVPEML -( .047) REVPCEM-( .047) VH HPPV -( .011) VHMPPV-I. ( .011) F CTYUTLN RV MPV I'I-( 056) (.011 ) RVHPVM-( .0 11) Fig. 4. 4 Weeke n d Serv i c e: D ec ision Hie r a r c h y and Wei ghts 43

PAGE 50

REVTVH( .011 ) OPCAPT(.01 3 ) JO EPI .O. (.007 ) REVENU E O E PHR ( 0 16) ( .009) GOAL --+OPOL VRY -OPCOST-J OEPPT ( ,2) ( .OSS) ( 0 06) PASSCST OEPPM (.013 ) ( 006) JMEPRML (.0 0 8 ) MANCST MEPO E -( 01 S) ( 007 ) JGASOP(.021) ENERGY ELEOP -( 0 4 6) ( .02S) FARE-( 042) NUSERSTP( 013) COMPI.S --t USERSAl'P ( 08 1 ) ( 0 36) SEATN"G ( 033) (.01 3) MAXACCPCONRELBUSESATR(28) (. 093 ) ( 035) RELBLTY (.025) MX.PACC ( 02) NUSCRSiT ( 018) \ M\NNRS6 (0 233) MWNR S 1 0 (0.230) UNSRaT6 (0.25 0) I UNSR 8T10 (0 2S8) Fi g. 4.4 W eekend Serv ice: Decisi on Hiera rchy and Weights (con t inued) 4 4

PAGE 51

,.,........ ..... RI'MOIIE -USCRSATT-(106) l (.05<4) TRVllt.E. -( .034 ) MAXEVNM l NOSPOLN -( 12) ( 0 54) A I RPO LN -(086) i AbbrtYUtion Otf ... itiOn I IACCDENTS 1 Number an
PAGE 52

Abbreviation UNSR8 T 10 U NSR8 T 6 MWNRS6 MWNRS10 Synthesis of Leaf Nodes with respec t to GOAL Distributive Mode Definition U n iform Service start 8 a m End 10 p m Unifrom Service Start 8 am, E n d 6 p m l Morn i ngs -P rayers Noon-Matts, Serv i ce end at 6 pm j Morni ng-Wo r sh i p, Noon-Shorp i r.gJRecreat ion, E n d 10 pm F ig. 4.5 Weekend Service: Evaluation of Alternatives 46

PAGE 53

I Crit% Air% .n .90 r .80 I '\ ' -' .30 ----. . UNSR8T10 .70 r-, \:' \/ "x . / U NSR8T6 .. 'J --4 .. .. . _,. ........... :\ ---hflo/NRS6 '>.,. -"-'. .Gn :', :\-,, ... .... ... 'l MWNRS10 .. . ... .. /; \ '. ..... . . .. '. . / .50 \ .20 . . . .. .. 0 .30 . 1 0 21l H .00 MAXPROD MAXACCP OVERAll 0 0 MAXSERV OPOLVRY MAXEVHM l Abbreviation Definition MA.XSERV l M txlmize Service M 0 0 f MIJUmize P roductivity OPOt.VRY J optimize means i!nd efflcitnct,2f sel'Vice de!Ntry MA.XACCP M.aDITize Pwon Acceptance MaJimze environmentl'd sounclne$s U NSR5T10 1 U niform Service stan 8 om, End t O p m UN SR5T 6 1 Uni! t o m SefVice Stan 8 am. End 6 pm MWNRS6 1 Morn1ngs-Prayers. Noon-Malls. SoMce end at 6 pm MWNRSIO J MomlngWors!Op. End 10 pm --Fig. 4.6 Weekend Service: Performance Sensitivity 47

PAGE 54

5. CONCLUSIONS ROI1da Statute 341.071 was enacted to require each provider to establiSh productiVity and pertormance measures for evaluating the effectiveness of their system. This StaMe provides an excellent motivation for providers to implement comprehensive data collecti on efforts to meet the data needs for their Transtt Deveopment Plans and for other needs, both adr!lnstrative and operational. lklfortunately, most of the IDP's reviewed minimally met the data needs. HeflCe, it was only possi ble to apply the product developed i n this study to a few case studies. lhese plans however, were the sources of the practical decision situations which were included i n Section 2 2 The set of objectives and measures of effectiveness which were used I n this study included environmental factors (noise and air pollution) and patron assessment of the system. This provides a comprehensive set of objectives and measures which cover sltuations expected to occur in decision mal
PAGE 55

6. REFERENCES Boyle, O.K. and P. E. Ouderkirk, Strategic Planning for Transtt Agencies in Small UrbaniZed Areas", Transportation Research Record 1402, TAB, National Research Council, Washington, D.C., 1993, pp. 25. Cherv.Qny, w. and M. C. Ferreri, "Bus Sketch PISnning", Transportation Research Record, Vol. 796, 1981, pp. 1 11. Cherwony, W. and M. C Ferreri, "Strategic Planning as a Transtt Management Tool", TranspOitation Research Record, Vol. 797, 1961, pp. 1-16. Collura, J and D. F Cope, "Assessing user Needs in Design of a Management Information System for Rural Public Transportation Services", Transportation Research Record 854, TAB, National Research Council, Washington, D.C., 1982, pp. 67. Collura, J. and P. McOwen, Management Information System for Small Fixed-Route, Fixed Schedule Operators, Transportation Research Record 994, TAB, National Research Council, Washington, D.C., 1984, pp. 71-75 Damm, 0., "InformatiOn Information Related Needs in the Transit Industry", Transportation Research Record 936, TAB, National Research Council, Washington, D.C., 1983, pp. 12-15. Davis, M w. Apptied Decision Support", Prentice Hall Book Co., 1968. El Sherif, H., M. D. Meyer adn N .H. M. WilSon, "Potential Role of Decision SuppOit Systems in Tranasit Managemnet", Transportalion Research Record, Vol857, 1982, pp 25-31. Fielding, G. J. and W. M. Lyons, "Performance Evaluation for DisCretionary Grant Transtt Programs", Transportation Research Record, Vol. 797, 1981, pp. 34-40. Fricker, J. D. and R. M. Shanteau, "lmproveel Service Strategies for Small-City Transit", TransportatiOn Research Record 1051, TAB, National Research Council, Washington, D.C., 1986, pp. 30. Guenthner R. P. and K. C. Sinha, "Transit Performance Evaluatioon Model", Transportation Engineering Journal, ASCE, Vol. 108, No. TE4, July 1982, pp.343. Holec, J. M. and R. L. Peskin, "Use ofProductivtty Measures in Projecting Bus and Rail Transit Operating Expenditures", Transportation Research.aecord, Vol. 797, pp. 40-49. Jensen-FISher, R., 'Transtt Needs for Planning Purposes", Transportation Research Circular, No. 407, April 1993, pp. 36-40. Kocur, G. and J. Tore, "FRACAS: A Strategic P l aming Model for Bus Transit Systems Transportation Research Record 994, TAB, National Research Council, Washington, D .C., 1984, pp. 13. Lutin, J. M., M. Uotile and T. M. Ash, "Modeing Transit Service Areas", Transportation Research Record, Vol. 797,1981, pp.16. MacDorman, L. C. "Extraboard Management Procedures and Tools", National Cooperative Transit Research & Development Program Report, June 1985, Section 5 : Syntllesls of Transtt Practice. Maze, T. H. adn U .outta, "Bus Maintenance Planning with Computer Simulation", Transportation 49

PAGE 56

Engineemg Journal, ASCE, Vol. 109, No.3 May 1983 ,pp. 389-402. Maze T. H., S Khasnabls, K Kapur and M.S. Paola, "Proposed ApproaCh 1o Oetemnlning Optimal 1\kJITtler, Size aoo Location of Bus Garage Additions, Transportation Research Record, Vol. 798, 1981, pp. 11-18. Pake, B E., M. J. Demetsky and L. A. Hoel, evaluation of Bus Maintenance Operations", Transpor1atlon Research Record 1019, TAB, National Research Council, Washington, D.C., 1985, pp. n-84. Santhakt.rrar, M and P. Harhlran, "Transportation Syatem Bus Management OptiOn! to lrrc>rove Urban Bus Route Performance Using Computer Simulation Transportation Research Record 1338, TAB National Research Council, Washington, D.C., 1992, pp. 22. Zaharia, T., Analysis of Urban Transpotatlon Criteria Transpof1artlon Engineering Journal, ASCE, Vol. 101, No. TE3, Proc. Paper 11496, August 1975, pp. 521. Zenilo, R. J., C. A. Keck and N R. Schneider, "Analysis of Transiot Performance Measures Used In New Vorl< State", Transportation Research Record, Vol. 797, 1981, pp. 52-58. Zerrillo, R. J., Use of Service Evaluation Plans to Analyze New York State Transit Systems, Transportatrion Research Record, Vol. 797,1981, pp. 58-61. 50

PAGE 57

APPENDIX I GENERAL INSTRUCTIONS FOR CREATING AN EC MooEL On invoking ECWIN (Expert ChOCe for Windows) the user is presented with a menu of icons. Of Interest in model building and utilization are the Structuring and the Evaluation and Choice Icons. A new model may be buitt starting from eijher icon. The general instructions for building starong from the Stnx:turing and the EvaluatiOn and Choice icons are itemized below as options A and B, respectively. A. :;!tructurioq Option 1. Double click on the Structurtng icon. 2. Select File then New. Enter goal descrtption 3. Type in a We name in which to store your new model. The extension .sm is added automatically by default. 4. Choose mode of structuring -top down (start w/ objectives), or bottom up (start w/ anernatives). We choose the latter in what follows. 5. Click the A button on the Icon bar to bring up the anematives window. 6. Type the description of the anernatives on the top column, press enter. Type the abbreviated name of the anernatlve in a smaller window that appears. 7. Repeat step 6 for all attematives. 8 Click the Structuring button on the Icon bar to bring up the objectives treeview window. 9. Type the description of the objective on the top COlumn, press enter. Type the abbreviated name of the anernative in a smaller window that appears, press enter 1 o Repeat step 9 tor all objectives, By pressing enter twice the user exits the Structuring option and iS returned to the Icon bar. The resulting model has an extension .STR I.e. TRANSIT.STR B. Using the .Ey.aluatioo.aod.Choice.Olltion 1. Double click on the Evaluation and Choice iC()A,2. Select File then New. 3. Type in a file name in which to store your new model. The extension *.EC* is added automatiCally by default. 4. Choose a mode to build the model in ( for now choose Direct to build the model in Evaluation and Choice directly) 5. Type in a description of the ggaldefinition (up to 65 characters). 6. After the goal node appears, select Insert from the Edtl menu to add the first "descendent" node and enter the node name in the box where the cursor appears. 7. Type In a description of the criterton (objective) the descendent node represents, and enter its name in the new box where the cursor appears. 8 Repeat these steps until all the peers at this level are entered. To finish entertng peers 51

PAGE 58

at a given level Simpl y press o r at the next blank node. Note that the decimal weights uneler the noele names change to maintain equal weights for the descendants with respect to the parent node each time you aelel a new peer. 9. To add another level of descendent noeles, select a node at the bottom of the tree under whiCh you want to add descenclants. Select Insert from tile Edi t menu anel repeat steps 7 and 8 above, adding peers and new descendants of the first level nodes as needed. If no further levels of descendants are required, then complete the structure of the moelel by entering the attematives (step #1 0). 10. The alternatives, a lso referred to as the "leaves", are entered in the same way as any of the descenelent noeles. If the same set of altematlves applies to any or all of the other corresponding parent nodes lhe entire set can be copied under individual selected options under the Repl i cate command inciiJCie replicating the current node' s children to Its peers, replicating the marked noele's children to the current node, and replicating lhe current node' s children to for m the leaves below all the other nodes at the bottom of the tree. Replicate copies only the children, not the descendants of the chi ldren nor the parents of the children. To copy children with all their descendants as well (i.e., a "plex), use the Copy Plex and Paste P t ax commands. 52