基于改进遗传算法的区域交通信号配时优化

Regional Traffic Signal Timing Optimization Based on Genetic Algorithm

针对城市网络的区域信号配时,本文建立了旨在最小化网络总延误的双层规划模型.在考虑出行者出行需求的基础上,以信号相位绿灯时长为控制变量,实现总延误最小化.在对用户出行需求的路径分配上,将流量分配模型转化为均衡路径问题,进而实现出行用户均衡.由于区域信号配时的变量随着网络规模的增加而增加,因此在求解多变量优化模型时,本文采用改进的遗传算法对该多变量优化问题进行分析和求解.以典型的城市区域交通网络为例,对该问题进行分析和算法的验证.算例表明,改进的遗传算法在城市区域网络中,能够有效地实现信号配时方案的优化,对于城市交通信号配时优化和管理有积极的启示.

This paper develops a bi-level model to optimize signal green timing and hence minimize the total travel delays under given urban area network. The network flow distribution related to traveler＇ s route choice with regards to their experienced delays is considered in the model with the variables of signal phase green time in every intersection. And then the flow distribution model is converted to the problem of travelers＇ best routes choice by the variational inequality theory, thus achieving stochastic user equilibrium. Because the number of variables increases with the scale of network, the improved genetic algorithm（GA） is presented to solve the multi-variable problem. The algorithm shows advantages in dealing with the problem and a typical urban area network example is analyzed. An example of typical regional signal optimization is presented to verify the effectiveness of the method. The result shows that the method can effectively create the optimal signal timing plans in urban areas.