Models And Algorithms On Improved Logit-based Stochastic User Equilibrium

With the overflying development of the economic in our nation,the income of our residents increases,whose car ownership is climbing sharply.Traffic problem has been the common concern of the government and ordinary citizens.To solve the traffic jam,the relevant administrators take the traffic supply and demand into consideration,increasing construction of traffic infrastructure and improving the level of traffic demand management simultaneously,to some extent,which brings success.Traffic assignment model is an important part of traffic planning and management and one of assisted means for making effective decisions.So,the paper mainly focuses on traffic assignment problem based on Logit models.The major studying aspects of the paper are listed as follows.(1)The theory of traffic assignment and discrete choice model applied in the traffic assignment is analyzed.Dial algorithm defines the effective routes of road network too strictly in the Logit-based stochastic assignment,leading to some routes with low impedance can't be used.To overcome the drawback,the paper redefines the effective routes and puts forward the route choice probability based on relative impedance.Meanwhile,a numerical example is given to prove the proposed algorithm.The results show that the proposed one has more reasonable results than the traditional one.(2)The paper expounds the form of route choice probability and the methods to determine the parameter in the path size correction Logit model.A path-size correction Logit-based stochastic user equilibrium model considering the length of different origins and destinations is proposed.The equivalence of the proposed model and the uniqueness of the solution are proved.The path-based successive averages(MSA)algorithm is designed to be solution algorithm.By the numerical example,the proposed model is compared to the polynomial Logit-based stochastic user equilibrium and the path-size correction Logit-based stochastic user equilibrium.The impacts of the changes in the new model on the flows of the network are analyzed.(3)Considering improved Logit-based stochastic user equilibrium model is usually solved by the method of successive averaging method,which has fast convergence rate at earlier stage,but slower later,the paper proposes a hybrid algorithm,which combines Tabu search algorithm with particle swarm optimization algorithm(TS-PSO)to solve the path-size correction Logit-based stochastic user equilibrium.And a numerical example is given,which is solved by the path-based MSA and TS-PSO.By the comparison and analysis of the results,the proposed algorithm is proved effective.(4)The travelers in the road network are not homogeneous,who have their ownpreference of the factors affecting the impedance.At the same time,the capacity of the sections in the network is uncertain.Under some stochastic factors,the travel time is uncertain too.The paper considers both expected travel time and the reliability of travel time,providing the generalized travel cost.On this basis,the paper presents multi-class path-size correction Logit-based stochastic user equilibrium model.The equivalence of the proposed model and the uniqueness of the solution are proved.The parameter sensitivity analysis is carried out on the given numerical example.The results show that the model is in accordance with the route choice behavior of the travelers.