Using decision field theory models in transport modelling: how far have we got and what can we do next? by Thomas Hancock

This presentation was delivered as part of the Institute for Transport Studies (ITS) research seminar series: https://environment.leeds.ac.uk/events/6/transport. Introduction by Professor Stephane Hess. Decision field theory (DFT), although popular in mathematical psychology, has only recently made the transition to choice modelling for consumer choices. One of the key differences that DFT has from standard choice models is that it has preference values for each alternative that update over time. This results in a different probability of picking each alternative depending on how long a decision-maker considers their alternatives. Theoretically, this provides a perfect model for explaining and modelling decisions where attributes of possible alternatives change (i.e, gap sizes in lane changing and gap acceptance choices). However, computational complexities of calculating the probability of alternatives in DFT models have resulted in simplifications. Typically, the number of deliberation timesteps has been set to a high value and thus the stochastic process is averaged out. This consequently ignores and fails to utilise the dynamic nature of DFT. This talk discusses the steps towards building a dynamic DFT. First, we demonstrate that DFT can be used effectively to model choices from typical SP and RP transport choice datasets. Second, we consider how the number of deliberation timesteps in a DFT model can be estimated as a function of choice response time. Finally, we discuss the remaining steps required to make DFT a properly dynamic model that can incorporate dynamic data.

Thomas Hancock, Research Student, ITS, University of Leeds

12/5/2018 1:53:43 PM