Dynamic traffic assignment model based on mixed-logit

Abstract

The inherent limitation and applicable condition of the Logit Model and Mixed Logit model are studied separately, parameters estimation and simulation analysis of mixed logit model are presented based on Monte Carlo. For network traffic assignment, a logit model is adopted to compute travelers' route choice probability at departure time. At nodes able to switch to another route, the mixed-logit model was presented to calculate travelers' switching probability. The total travel time and route choice probability were compared between two conditions of considering and non-considering switching routes. At last, an example is shown to test the model and algorithm.