基于改进NSGA-Ⅱ的交叉口信号配时多目标优化
Multi-objective optimization of intersection signal timing based on improved NSGA-Ⅱ algorithm
陈秀锋 1王瑞聪 1陈咨羽 1高梦圆 1郭玉彤1
作者信息
- 1. 青岛理工大学 土木工程学院, 青岛 266525
- 折叠
摘要
针对城市交叉口信号配时多目标优化的求解效率问题,设计了一种基于改进NSGA-Ⅱ的城市交叉口信号配时优化方案.该方法采纳车辆延误、停车次数、通行能力3个评价指标设计了多目标信号配时优化模型,并提出了改进的NSGA-Ⅱ(NSGA-Ⅱ-DE)对配时模型进行求解.NSGA-Ⅱ-DE采用DE/Rand/1的全局变异策略对NSGA-Ⅱ算法选择策略进行改进,并设计了种群动态更新算法,提高了算法的收敛速度和求解精度.结果表明改进算法获得了较好的Pareto最优解,通过典型交叉口仿真验证,相对于NSGA-Ⅱ算法,NS-GA-Ⅱ-DE方法设计配时方案平均车辆延误优化提升13.42%,排队长度优化提升16.73%,平均停车次数优化提升15.50%,能够实现城市交叉口的最优化控制.
Abstract
A method for optimizing signal timing at urban intersections based on the im-proved NSGA-Ⅱ algorithm is proposed in this study to address the efficiency problem of multi-objective optimization of signal timing at urban intersections.The method first builds a multi-objective signal timing optimization model by using the three performance indicators of vehicle delay,number of stops and capacity,and then proposes an improved NSGA-Ⅱ(NSGA-Ⅱ-DE)algorithm to solve the timing model.The NSGA-Ⅱ-DE modifies the NSGA-Ⅱ algorithm selection strategy by using the global variation strategy of DE/Rand/1,and in-creases the convergence speed and solution accuracy of the algorithm by designing the popu-lation dynamic update algorithm.The results show that the improved algorithm obtains a better Pareto optimal solution.Through the simulation verification of typical intersections,it is found that compared with the NSGA-Ⅱ algorithm,the average vehicle delay optimiza-tion of the timing scheme of this method is improved by 14.04%,the queue length optimiza-tion is improved by 18.39%,and the average stop number optimization is improved by 18.22%.The optimization method proposed in this study can realize the optimal control of urban intersections.
关键词
交通工程/信号配时/多目标优化/T形交叉口/城市交通/相位/NSGA-Ⅱ-DE算法/全局变异Key words
traffic engineering/signal timing/multi-objective optimization/T-shaped intersec-tion/urban traffic/phase/NSGA-Ⅱ-DE algorithm/global variance引用本文复制引用
基金项目
国家自然科学基金(62003182)
山东省自然科学基金(ZR2019MG012)
山东省自然科学基金(ZR2020MG021)
山东省重点研发计划(2019GGX101038)
出版年
2024
NETL
NSTL
维普
万方数据