Dynamic traffic assignment models of route and departure time choice for travelers through congested networks provide important insight into the dynamics of peak periods and sensitivity of travellers' behaviour in response to a range of transport policy measures. This study aims to analyse the dynamic system optimal assignment (DSOA) with departure time choice, whichis an important, yet underdeveloped area. The DSOA process suggests that there is a central 'system manager' to distribute network traffic over timewithin a fixed horizon. Consequently the total, rather than individual, travel cost of all travelers through the network is minimised. Although thesystem optimal assignment is not a realistic representation of network traffic, it provides a bound on how the best use may be made of the road system, and as such it is a useful benchmark for evaluating various transportpolicy measures. The formulation and necessary conditions of the DSOA arederived. The DSOA problem is formulated as a state-dependent optimal control problem. To analyse the dynamic externalities, a novel sensitivity analysis of the traffic model with respect to the link inflow is adopted. With the solution algorithms, some numerical calculations are provided and the characteristics of the results are discussed. For the covering abstractsee ITRD E144727. Reprinted by permission of Elsevier.
Abstract