The development of improved understanding of the dynamical properties of transport systems is widely recognised as being of vital importance in a wide range of different theoretical and practical transport modelling contexts. Amongst the many behaviour factors underlying the dynamic properties of transport systems, the mechanisms of traveller learning are to date perhaps the most poorly understood. In particular, very little is known about the influence of travellers' learning mechanisms on the responsiveness and evolution of transport systems, over various time scales. The objective of this paper is to investigate the influence of a range of different learning mechanisms on the dynamics of transport systems. This problem is studied in the context of stochastic traffic assignment, which provides a natural framework for investigating day-to-day learning processes. The research makes use of a recently developed stochastic assignment procedure that admits both within-day and between-day dynamics.
Abstract