Research On Combinatorial Optimization Of Signal Phase Design And Signal Timing For An Isolated Intersection

Intersection signal phase design and signal timing are mutually restricted and coupled.However,the previous researches on signal control at an isolated intersection mostly separated the two and ignored their internal connection,which to some extent affected the effect of signal control.Therefore,it is theoretically and practically significant to study combinatorial optimization of signal phase design and signal timing for an isolated intersection.First,this paper provides a review of the previous literature on signal control at an isolated intersection,and then introduces the related concepts,typical control methods and optimization algorithms.Second,in view of the associsted existing deficiency,this paper formulates a combinatorial optimization model with respect to signal phase design and signal timing to minimize the average individual vehicle delay of an isolated intersection.This model is a typical NP-hard problem,and the solution efficiency of brute force search and general intelligent algorithm is not high.Thus,an Elite Multi-population Genetic Particle Swarm Optimization(EMGA-PSO)is proposed,which integrates genetic algorithm,particle swarm algorithm and dynamic programming algorithm.It uses population diversity operator and elite retention strategy to connect and feedback the upper and lower layers of the algorithm,and consequently solves the combinatorial optimization model.In order to further compare and analyze the optimization effectiveness of our proposed combinatorial model and its algorithm,this paper proposes a two-stage optimization model and the associated algorithm.The signal phase design is optimized to minimize the total flow ratio at the intersection in the first stage.In the second stage,the signal timing is optimized to minimize the average individual vehicle delay,based on the signal phase design scheme obtained before.Finally,the intersection of Zhongxing South Road and Chengnan Avenue in Shaoxing City,Zhejiang Province,is taken as an example,and the signal control effectiveness of the TRRL method,two-stage method and combinatorial optimization method are analyzed and compared with the current signal control.Numerical results show that compared with the current signal control,the combinatorial optimization method,two-stage method and TRRL method reduces the average vehicle delay by20.2%,4.3%,3.3% respectively,which verifies the effectiveness of the combinatorial optimization.In addition,the simulation results also shows that it can improve the effect of signal control.