A Link-based Traffic Assignment Model Based On Dynamic Discrete Selection

With the growing contradiction between the supply and demand of urban traffic in China,the traffic problems have become one of the major bottlenecks which restrict the urban development.The conventional ideas that increasing investment to traffic supply can hardly solve this problem radically.Under the constraint of land resources,planning and managing roads correctly is no doubtly the most effective way.The traffic assignment problem(TAP)is one of the most significant footstones of the modern theory of urban traffic planning.It is also a key technology of the urban traffic network design.In this circumstance,this paper aims to formulate a model for traffic assignment problem which can describe the travelers’ choosing behavior more accurately.By doing this,the link-based flow pattern can be obtained and be used by traffic managers and planners for decision making.Based on the analysis of the charicteristics of the network and the perception abilities to network of the travelers,this paper proposes the viewpoint that travelers are always myopic in the process of making decisions.The author also defines the concept of myopic decision and myopic behavior in transportation field.Based on the theory of dynamic descrete choice,this paper establishes a link-based myopic RL model which takes travelers’ myopic behavior and node cost into consideration.The myopic RL model substitutes path choice with a sequence of link choices.The myopic parameter is introduced to avoid the assumption of drivers usually acquiring spatial knowledge of networks through direct environmental experiences,and to describe the chooing behavior under virious circumstance more generally.Moreover,there is no need to establish the path set in myopic RL model.By setting the random term independently and identically distributed with the Gumbel distribution,the author formulates the expression of myopic RL model and designs an algotrithm for solving this problem.In this paper,the author takes a small-scale network as an example,and compares the calculated results of myopic RL model,RL model,link-sized RL model,MNL model and PSL model.After analyzing the differences and connections among myopic RL model,MNL model and RLmodel in detail,the advantages of myopic RL model is illustrated.This paper analyzes the existing SUE models and gives out the mathematical programming of link-based traffic assignment problem based on myopic RL model.By comparing the existing link-based algorithms,the author decides to use the MSA algorithm to solve the problem.A small-scale simple network,a small-scale complex network and a middle-scale complex network are used as examples to illustrate the TAP based on myopic RL model,RL model and MNL model respectively.The calculated results are illustrated with comparative analysis.