Global Optimal Routing Algorithm for Traffic Systems with Multiple ODs

Abstract

Global optimal routing for multiple Origin-Destinations (ODs) in traffic systems becomes extremely complicated when considering the traffic volumes on the road sections. This paper proposes a Combinational Algorithm which is combined by Conventional Method, U Algorithm, SU Algorithm, SRU Algorithm, SAU Algorithm and SRAU Algorithm to solve this problem. Among the above 6 algorithms, SRAU Algorithm contributes to the Combinational Algorithm the most, where firstly, all original ODs are sorted by their traffic volumes, and then the order is randomized to generate some routing candidates. For each candidate, before finding the optimal route of the current OD, the traveling times on the optimal routes calculated by ODs with high priority are adjusted and then Q Value-based dynamic programming is utilized to find the optimal route. Next, an updating process is needed to update the traveling time on the route using the current OD. Finally the best solution can be selected out of all solutions. Sufficient simulations show that the proposed routing algorithm is efficient enough to obtain the near optimal solution even in very large scale traffic systems. Also the consideration of the traffic volumes on the road sections enables our proposal to apply to real traffic systems.