Transport Modelling in a Few Hours at Imperial College
For nearly ten years OmniTRANS has been an element of the Masters in Transport course at the Centre for Transport Studies run jointly by Imperial College London and University College London.
The Master course is broadly based and offers the students a wide range of modules and options, which translates into a very tight timetable. A general introduction to transport modelling is provided, but students who are interested need to select the Advanced Modelling module. Much of the course, naturally, is concerned with modelling theory, but attention is given to linking this with practical transport modelling, which is where OmniTRANS becomes relevant. Last year changes were made to align student projects involving the application of OmniTRANS with the course’s elements on discrete choice modelling (DCM) and network equilibrium.
The use of OmniTRANS on the course is based on a dataset known as ‘Studytown’. This has just 25 zones to enable students ready and free access to the OmniTRANS software, but the highway and public transport networks are based on adapted real data and provide considerable detail. There is also detailed information on demographic and land use information in the system. The recent changes added ‘personal’ data in the form of survey information about people’s travel choices.
This ‘survey’ needed to be synthesised, of course, for the fictional study area. This required modelling in OmniTRANS to identify travel costs by different modes and then, using appropriate sampling, to allocate ‘observed’ mode choices for car, public transport, and walking or cycling, taking account of income levels, car and season ticket availability for the origin zone of the trips, as well as the gender of the traveller. The availability of season tickets depended on the number of base year public transport trips from each zone.
The students’ projects therefore involved calibrating a discrete choice model using the travel survey dataset, which had some 300 survey ‘responses’, and then, using the calibrated parameter values within a choice model implemented in OmniTRANS (via OtChoice), to calculate changes in travel demand for two scenarios that limited car traffic, firstly by applying financial charges to using the city centre part of Studytown and, secondly, by making the centre pedestrianised.
As the time available for the project work could be measured in hours, this provided a considerable challenge for the students.
Using a workbook developed for the students and exploiting facilities in OmniTRANS, the Studytown modelling system was arranged to allow students to concentrate on the underlying concepts, and avoid being overwhelmed by the scope of OmniTRANS’s capabilities. Nevertheless, students had to operate the user interface, edit networks, cope with Ruby scripts, and present the results.
Their work is shown in the following illustrations that indicate:
- Understanding the characteristics of the Studytown Area (in terms of existing mode choices)
- Running Biogeme calibration software with the survey data and cost information generated from OmniTRANS
- Adding the calibrated parameters to other modelling parameters via the OmniTRANS Parameter Manager
- Running the choice modelling and viewing the effects on public transport and walk modes
- Comparing results on the highways for different scenarios.
The students made presentations and submitted reports from their project work, from some of the images are taken.
