基于Pareto粒子群算法的路口多目标信号控制模型  

The multi-objective signal control model for intersections based on particle swarm algorithm with Pareto optimal solution set

作　　者：李巧茹 李欣 陈亮 LI Qiaoru;LI Xin;CHEN Liang(College of Civil Engineering,Hebei University of Technology,Tianjin 300401,China;Civil Engineering Technology Research Center of Hebei Province,Tianjin 300401,China)

机构地区：[1]河北工业大学土木与交通学院,天津300401 [2]河北省土木工程技术研究中心,天津300401

出　　处：《铁道科学与工程学报》2018年第4期1076-1087,共12页Journal of Railway Science and Engineering

基　　金：北京城市交通协同创新中心资助项目;河北省高等学校科学技术研究重点资助项目(ZD 2014078)

摘　　要：为提高路口的运行效率,实现交叉口信号配时的实时动态调整,基于Pareto最优化多目标粒子群算法,建立延误和停车次数最小、有效通行能力最大的路口多目标信号控制模型。由于各个评价指标之间相互冲突且量纲不同,属于非劣问题,分别比较不同评价指标得到多目标信号配时的非劣解,更接近最优解。因此,模型根据Pareto支配关系与密度距离进行粒子选择,最终得到路口信号配时模型的Pareto最优解。研究结果表明:该信号配时模型所得到的评价指标优于路口现状配时以及基于单目标最优化的信号控制模型,能够应用于实时的路口信号控制。In order to improve the operation efficiency of intersections and realize the dynamic characteristics adjustment of signal control parameters,a multi-objective signal control model for intersections was established based on particle swarm algorithm with Pareto optimal solution set.The delay and stop rate were minimized,while the effective traffic capacity was maximized.However,the multi-objective model was a non-inferior problem and must be solved by comparing every evaluation index separately because each evaluation index changes respectively and has different dimension.This model applied particle swarm algorithm based on Pareto optimal solution set which obtains the optimal signal timing solution by utilizing Pareto non-dominated relation and density distance to order the particles and choose the better particles.The case study shows that this model can gain the preferable signal control scheme on performance of traffic evaluation indices compared with current fixed signal control method and single objective signal control model converted from multi-objective model.It can be applied to real-time signal control intersections.

关 键 词：交通工程 多目标信号控制模型 Pareto最优化 路口信号配时 粒子群算法 

分 类 号：U491.23[交通运输工程—交通运输规划与管理]