A Semi-Dynamic Assignment Model Considering Drivers' Responses to Travel Time Uncertainty

Abstract

In this paper, we assume that network flow is normally distributed due to stochastic travel demands, and develop a stochastic model of semi-dynamic assignment that is able to consider congestion's time-space movement. Also, drivers' risk attitude to uncertain travel time is incorporated using the efficient travel time.We formulate the stochastic model of semi-dynamic assignment as a non-linear complementarity problem, and propose an algorithm with the conjugate gradient method. Further, the model is applied to Kanazawa road network and examine the validity of the model and applicability to a real-world network