Apivat Jotisankasa
Imperial College London
, London, United Kingdom, John Polak, Centre for Transport Studies
Imperial College London
, London, United Kingdom
Day-to-Day Traveller Learning and Adaptation in Route and Departure Time Choice Behaviour (assigned to theme
Understanding the dynamics of individual travel choice behaviour is widely regarded as one of the most challenging topics in travel behaviour research. Of particular interest are the day-to-day dynamics of commuters’ route and departure time choice decisions as these decisions determine the major flow patterns in transport networks during peak periods. Two of the key mechanisms underlying these behavioural dynamics are the process by which commuters learn about travel conditions and the process by which they adjust their behaviour as a result of new experiences on the networks. A number of recent simulation studies using different learning and behavioural adjustment mechanisms have highlighted the fact that these processes can have a profound effect upon the dynamic properties of transport systems, both in terms of the temporal evolution and the final states of the systems, and thus upon the evaluation of transport policies. In the view of the importance of this effect, it is surprising that little work has been carried out to establish the empirical properties of these mechanisms. This paper presents a new theoretical framework for modelling the day-to-day dynamics of drivers’ route and departure time choice decisions in a commuting context. The proposed framework consists of two inter-related components; the first one being a travel time perception updating process and the second being a choice process, which si conditional on current perceptions. The resulting sequential decision process is formulated as a dynamic discrete choice model using a mixed multinomial logit form. Data for the estimation of this model are provided by a series of web-based experiments involving actual travellers making repeated decisions in hypothetical transport networks over a period of 15 simulated days.
 | Conference organized through conf-vienna
(copyright Gunther Maier) |