evolution of the greater vancouver transportation model.docx

1、evolution of the greater vancouver transportation modelEvolution of the Greater Vancouver Transportation Model13th Annual International EMME/2 Users ConferenceHouston, TexasOctober 28-30, 1998by Karoly KrajczarGreater Vancouver Regional DistrictBurnaby, B.C.1. IntroductionThis paper provides an over

2、view of the evolution of the Greater Vancouver Regional District (GVRD) Transportation Model and focuses on the recent enhancements that were completed in 1998 as part of the GVRDs Transportation Model Enhancement Program. This paper describes how the improvements were identified and prioritized and

3、 provides a summary of the model enhancements. The other papers in this session will elaborate on three of the major enhancements (Transit captivity, HOV, Park and Ride). 2. Evolution of the Greater Vancouver Transportation ModelThe 1970sThe GVRD has a long history in the development and application

4、 of transportation models dating back to the mid-1970s. In 1975, the GVRDs first computer-based transportation model was developed in conjunction with the University of British Columbia (UBC) Computer Science Department. The model consisted of 125 traffic zones, a road and transit network and algori

5、thms to implement a traditional four-step transportation planning model for the morning peak period. Specialized software was written to implement the model on the UBC mainframe computer. The model had limited graphical capabilities, with output consisting primarily of tables of link volumes and oth

6、er attribute data.In 1979, the model underwent a significant re-calibration in preparation for the GVRD Rapid Transit Project which was a comprehensive set of studies of future rapid transit lines and technology options for the GVRD. At this time, the traffic zone system was expanded to 152 zones, t

7、he transit assignment algorithm was enhanced and the four-step model was re-calibrated. The model was calibrated using travel data from the 1978 North Vancouver Origin-Destination Survey and a series of studies by the City of Vancouver including the 1975 Downtown Vancouver OD Survey, the 1976 Transi

8、t Study and the 1977 Cordon Survey.The 1980sIn 1983, the GVRD conducted a comparative review of transportation modelling software. The GVRDs existing software (developed in-house) was compared with the Federal Highway Administrations Urban Transportation Planning System (UTPS) and INRO Consultants E

9、MME/2. The review committee selected EMME/2 and the GVRD became the second North American site to obtain EMME/2 (Portland, Oregon was the first site).In 1984 the model was re-calibrated in conjunction with the implementation of EMME/2. The traffic zone system was expanded to 370 zones, road and tran

10、sit networks were updated and the four-step model was calibrated to 1981 Census data which contained the place of work-place of residence survey. Significant changes to the model structure included the introduction of time-space equations for auto work trips and combined impedance for work trip dist

11、ribution. At this time, the GVRD purchased the Pixel computer (a 32-bit micro computer) to reduce the high computing costs associated with running the model at UBC. In those days the macro language was not available in EMME/2 and the model equations were entered directly into the matrix calculator (

12、each matrix calculation took about 30 minutes and a 15 iteration assignment required approximately 14 hours of computing time).In 1985, a comprehensive set of 24 hour travel surveys was conducted to provide a description of travel patterns in the GVRD prior to SkyTrain. The project included a 25,000

13、 household telephone interview survey (a five percent sample of the 522,000 households in the Vancouver CMA), a 1,700 household trip diary survey (a 0.3 percent sample), traffic screenline counts, an external cordon survey and a taxi and business travel survey. Data from the household survey is cont

14、ained in a series of three reports.Between 1986 and 1987, the model was re-calibrated to the 1985 survey data. The traffic zone system was increased slightly to 380 zones, networks were updated and the four step model was re-calibrated. The model was then validated to post-Skytrain conditions in 198

15、7. Although several major new facilities had been introduced to the region between 1985 and 1987 (e.g. Skytrain, Alex Fraser Bridge, the East-West Freeway in Richmond), the model compared closely with 1987 screenline counts.One of the first applications of the new regional model was to examine the i

16、mpacts of long-term growth on major transportation facilities by the City of Richmond. The findings of this study showed that with current trends land use many of the major river crossings would be at capacity by 1996. This study resulted in an initiative by Richmond council to request that the prov

17、ince lead a comprehensive inter-agency study on land use and transportation for the GVRD. At the same time, many of the GVRD municipalities began to develop their own sub-area transportation models that contained more detailed zone systems and networks to deal with municipal transportation planning.

18、 Burnaby, Surrey, Coquitlam, Richmond and North Vancouver were some of the first municipalities with their own sub-area models. Many of these models were afternoon peak period models as opposed to the GVRD morning peak period model. The sub-area models were run by dialing-in to the GVRD Pixel comput

19、er.In 1988, the City of Vancouver and GVRD conducted a regional trucking survey. Information from this survey was used to calibrate a 24 hour truck model. This model provided information on two categories of truck movements - light trucks with a GVW of 4,500-20,000 kg and heavy trucks with a GVW ove

20、r 20,000 kg.Due to increased demands on EMME/2, a decision was made in 1989 to transfer from the Pixel computer to the GVRD VAX computer. A multiple EMME/2 VAX license was purchased from INRO Consultants Inc. which provided municipalities and government agencies with better access to their models. T

21、he result was a significant increase in speed and expanded use by various agencies.A major 1989 study to make extensive use of the regional model was the Freedom to Move - Greater Vancouver Transportation Task Force study. This study, which was sponsored by the Provincial Government and the GVRD, pr

22、oduced an integrated transportation plan for the GVRD to the year 2001.The 1990sIn 1991, the regional model and truck model were validated to current conditions. The validation process included expanding the traffic zone system to 445 zones, updating the land use inputs and networks and some modific

23、ations to the four-step model and significant changes to the truck model structure. During the validation exercise, the GVRD decided to transfer EMME/2 from the VAX to the PC. Again, the reason being lower costs, better access to the models and faster computing time for users.This work was done in p

24、reparation for Transport 2021 which was the regions first long-range transportation plan with associated policies on land use, transportation demand management (TDM) and investment priorities for transit and road. Transport 2021 was a two year project that made extensive use of the regional model fo

25、r testing land use scenarios, TDM strategies and road and transit networks.In 1992 the GVRD, BC Transit and the Ministry of Transportation and Highways jointly funded a major household travel survey for the morning peak period (6:00 - 9:00 a.m.). This telephone survey collected data from 15,000 hous

26、eholds (a two percent sample of the 690,000 households in the Lower Mainland - Lions Bay to Chilliwack) with information on more than 25,000 morning trips. One of the primary objectives of this survey was to collect data required to re-calibrate the regional model to current morning conditions. This

27、 was followed by a 24 hour trip diary survey in 1994. This was a mail-back survey that collected information from 1,600 households (0.2 percent sample of the 740,000 households in the Lower Mainland) with information on more than 11,000 trips throughout the day. The trip diary was designed to collec

28、t baseline information for regional planning purposes, as well as for calibrating an afternoon peak period transportation model.In late 1994, the Province, GVRD and BC Transit embarked on a follow-up study to Transport 2021 called the South Coast Transportation System Plan. This study was intended t

29、o provide a more detailed description of the infrastructure requirements identified in Transport 2021. At about the same time, BC Transit was involved in the development of their 10 Year Plan and in a Multiple Account Evaluation of Rapid Transit Options for the GVRD. Both of these studies made exten

30、sive use of the regional model and the GVRD truck model.In 1995, the GVRD conducted a re-calibration of the model to the 1992 Travel Survey. In summary, the traffic zone system was expanded from 445 zones to 652 zones, the road network was expanded from 7,400 to 13,000 links, the transit network was

31、 updated and new intersection-based volume-delay functions were introduced, replacing the old Bureau of Public Roads (BPR) functions.In 1996, the GVRD established the Transportation Model Enhancement Program to address the evolving needs of the model-users in the region. The first step involved upda

32、ting the model to 1996 conditions, developing a preliminary PM Transportation Model, collecting additional data and identifying future model enhancements. The identification of future model enhancements involved input from consultants, municipal and provincial staff. This feedback was categorized an

33、d prioritized (depending upon need, data availability, difficulty, cost, inter-dependencies, etc.). The result was a short, medium and long-term modelling research program that focuses on model enhancement, model expansion, user support and data collection.The short-term work program was completed in 1998 and invol