Abstract:Aiming at the traffic congestion problem of small and medium traffic intersections in cities, this paper proposed a variable phase sequence signal control model based on Timed Petri Net(TdPN). This paper used TdPN to establish the intersection traffic model and signal control model and also used Markov chain to establish the dynamic generation model of traffic flow. This paper assign a grant to the phase which had the largest number of waiting vehicles to select phase randomly. With the minimum average delay time as the optimization goal, this paper used genetic algorithm to solve the optimal phase timing. In the case of fixed signal period, this paper analysis the influence of TdPN-based four-phase variable phase sequence control model on the average queue length of intersections under unbalanced traffic flow. The simulation compared this with the four-phase fixed phase sequence control model. The research results show the scheme reduced the average queue length of intersections per unit time.