Abstract
A new approach is proposed for the analysis of spatial-interaction behavior under conditions of congestion. This approach, which is based on a cost-efficiency principle of interaction behavior, is shown to lead to a new class of exponential probability models for realized spatial-interaction patterns. In addition, these exponential trip-pattern models are shown to yield a natural stochastic generalization of the deterministic ‘user-equilibrium’ models currently employed in the analysis of urban traffic flows. In particular, such user-equilibria now constitute the ‘most probable states' in these new models.

This publication has 7 references indexed in Scilit: