基于成对可替代子路径的交通分配改进算法

摘要

This paper studies traffic assignment by paired alternative segments,which was designed by Bar-Gera.TAPAS focuses on acquiring the entropy user-equilibrium route flow vector,which needs to store vast PASs and then select some to shift flows.Since many traffic assignment algorithms concentrate on increasing the rate of convergence and simplification,this paper focuses on building a new effective PAS,shifting flows between segments of an effective PAS in order to balance the costs and redistribution of PAS flows between origins accord-ing to the condition of proportionality.The management of PAS set,including storage and elimination,are aban-doned.Three sub-algorithms are designed,including building a new effective PAS,shifting flows between seg-ments of an effective PAS and redistribution of PAS flows between origins.Numerical results were presented for a test network.The results show that,for some cases,the rate of convergence of the modified algorithm could be improved and the algorithm structure is much simpler,so the programming gets easier in the process of applica-tion.%以Bar-Gera设计的基于成对可替代子路径（PAS）的交通分配算法（TAPAS）为研究对象，注意到TAPAS算法为了追求最大熵用户均衡交通分配，存储了大量PAS并从中选择PAS集合进行流量转移。由于很多交通分配问题更加专注于提高算法的收敛速度和其简易性，将算法的核心操作仅限于当前新构建的有效PAS（不再存储PAS集合），在新构建的PAS上转移流量以均衡子路径费用，在相关起点之间均衡路径流量比例。设计包括有效PAS构建子算法、基于PAS的流量转移子算法和基于PAS在相关起点之间均衡路径流量等比例分配子算法。通过数值试验算例分析，结果表明：相比原算法，仅基于当前新构建PAS上进行流量转移的交通分配改进算法的收敛效率不会降低，同时，改进后的算法结构更简单，编程工作量更小，对部分案例来说更适于应用。