基于聚类思想的交通信号相位组合优化研究

为合理地进行信号相位组合的优化设计，从单点信号控制交叉口车辆运行特性出发，对相位组合优化问题进行分析。首先定义相位绿灯时间的相对利用率，在此基础之上，将相位内交通流的组合优化问题抽象为不同交通流对象之间的聚类问题，运用聚类分析方法建立相位组合优化模型。模型选取相位绿灯时间的相对利用率作为聚类结果的检验标准，将交叉口延误时间作为相位组合优化方案的评价指标。通过vissim仿真对比，信号交叉口总延误降低15%左右。

Signal Phase Composition Optimization Based on Cluster Algorithm

In order to design the signal phase composition rationally, starting with the figure of vehicle operating characteristics, we defines the relative utilization of phase green time by analyzing the phase combinations. Thus, the problem of traffic composition is viewed as clustering problem, and a phase combinatorial optimization model is constructed.By selecting relative utilization of phase green time and delay time as evaluation index of phase combinatorial optimization, through the comparative analysis of vissim simulation, total delay of signalized intersection is reduced by about 15%. This provides a new method for signal phase structure design.