Statistical Inference for Multi-Species ZRP Based on Variational Bayes Method

Abstract

Abstract. Traffic flow models with cellular automata have been applied to analysis of traffic phenomena to understand traffic jams and to construct urban traffic simulators. Estimating vehicular characteristics from the traffic flow is expected to realize the high performance and reproducible simulation. In this paper, we propose an estimation method of the parameter of a multi-species ZRP based on the variational Bayes method, and detect the OV function, which shows the optimal speed with respect to the distance between cars, from the spatio-temporal diagram.