A Quasi-dynamic Traffic Assignment Model with Stochastic User Equilibrium

Dynamic traffic assignment is a hot topic and difficult issue in the transportation research field, where the application of dynamic traffic assignment methods to large-scale networks often results in high computation costs. By studying the quasi-dynamic traffic assignment problem, the time continuity of dynamic traffic assignment and the complexity of corresponding models can be reduced. A new calculation method of residual traffic demand is defined based on link travel time. Then, considering the propagation of residual traffic demand between time periods and the differences in the familiarity of travelers with the road network, a Logit-based quasi-dynamic traffic assignment model with stochastic user equilibrium is established. An algorithm is designed to solve the model based on the successive average method. Finally, the application of the model is conducted on a Braess network and a nine-node network. The sensitivity analysis of model parameters is carried out to verify the effectiveness of the model and the algorithm. Results enrich the basic theory of the transportation research field and can provide references for transportation policymakers.