Coordinated Control Method Of Urban Traffic Signal And Guidance Under Stochastic Conditions

Traffic signal control and traffic guidance are effective approaches to deal with urban traffic congestion.The coordinated control of the two can improve the efficiency of the urban transportation system,which is a research hotspot in the field of traffic management and control.The bi-level model is one of the classic frameworks to deal with the coordinated control problem of the traffic signal and guidance.The upper model is usually the traffic signal control model that implements traffic signal timing optimization,and the lower model is usually the traffic assignment model that implements traffic guidance.At present,there are few researches on the traffic supply stochasticity and time-varying traffic demand stochasticity in the research of coordinated control of traffic signal and guidance based on the bi-level framework.In addition,in the lower traffic assignment model,the route choice model rarely considers the overlapping path problem and perfectly rational problem of travelers at the same time,which makes it difficult to accurately describe the actual route choice behavior of travelers,reducing the effect of traffic assignment.Therefore,in view of the above problems,the following aspects of research work are developed in this paper:First,a bi-level model framework is used to establish the bi-level coordinated control model of traffic signal and guidance under the stochastic supply and stochastic demand conditions.Among them,the upper model is a traffic signal control model.The lower model is a dynamic traffic guidance model,that is,the reliability-based dynamic traffic assignment(RDTA)model.In this bi-level model,the time-varying probability distribution characteristics(mean and variance)of origin destination(OD)demand are used to capture the stochasticity of traffic demand,and the degradation characteristics of link capacity are applied to reflect the stochasticity of traffic supply.In the interaction relationship between the upper model and the lower model,the green time calculated by the traffic signal control model will be transferred to the R-DTA model,and the time-varying mean path flow calculated by the R-DTA model will be transferred to the traffic signal control model.Second,this thesis proposes a route choice model considering the overlapping path problem and perfectly rational problem of travelers under the stochastic supply and stochastic demand conditions.In the above bi-level model,the route choice model is the key component of the lower R-DTA model.In this thesis,generalized nested logit(GNL)model is used to solve the overlapping path problem,and cumulative prospect value(CPV)is used to solve perfectly rational problem of travelers.In the GNL model,CPV is used to replace the expected utility value(travel time)to establish a CPV based GNL route choice model,which can simultaneously deal with the overlapping path problem and perfectly rational problem of travelers.It should be noted that the path travel time budget is determined by considering intersection delay based on stochastic link capacity and time-varying stochastic OD demand.The minimum path travel time budget is taken as the reference point for travelers’route choice.and combined with the value function and probability weight function,the path cumulative prospect value is calculated.Third,considering the actual requirements of coordinated control of dynamic traffic signal and guidance,a dynamic coordinated control approach under the stochastic supply and stochastic demand conditions is established based on the bi-level coordinated control model.The approach mainly includes four steps:data sorting,path flow calculation,traffic signal timing optimization and result output.More specifically,in the data sorting step,the input parameters and output parameters of the bi-level coordinated control model are given.In the path flow calculation step,the time-varying probability distribution characteristics of OD demand are input into the lower R-DTA model and solved by the method of successive averages(MSA)to obtain the time-varying mean path flow.In the traffic signal timing optimization step,the time-varying mean path flow is input into the upper traffic signal control model,and genetic algorithm(GA)is used to solve it to obtain the best green light time of the current generation.Then,the green time is used to calculate the intersection delay,which is transferred to the next iteration R-DTA model.In the result output step,the results of optimal signal timing and traffic equilibrium assignment satisfying the maximum evolution algebra are output.Finally,the Braess network and Nguyen-Dupuis network are applied as test networks,and numerical examples are given.To this end,for these two networks,an OD demand matrix with time-varying stochastic characteristics is constructed.Considering the link capacity degradation coefficient,the overlapping path problem,perfectly rational problem of travelers,the performance of R-DTA model and the performance of the coordinated control method of traffic signal and guidance under the stochastic supply and stochastic demand conditions are analyzed.The results indicate that the coordinated control approach can deal with the overlapping path problem and perfectly rational problem of travelers and reduce the total delay of the road network.The effectiveness of the model and method is verified.