摘要
从城市交通控制目标多样性的本质出发,考虑机动车时间效益、行人时间效益及环境效益,建立以机动车、行人流量为输入,机动车延误最小、行人延误最小及机动车停车率最小为目标的信号控制交叉口周期时长多目标优化模型,简称MOCLO模型.并应用多目标连续蚁群优化算法求解.算例的求解结果显示,连续蚁群优化算法能够均匀地逼近MOCLO模型的Pareto最优前沿的各部分;与F-B方法、ARRB方法相比,MOCLO模型对周期时长的优化结果在机动车时间效益、行人时间效益及环境效益三方面的综合性指标较好;MOCLO模型可提供多个不同特性周期时长以满足不同交通状态的需求.
In view of the multi-objective property of urban traffic signal control, the pater presents a multi-objective cycle length optimization（MOCLO）model with the vehicle volume and the pedestrian number as the inputs and minimum vehicle, pedestrian delay and stops as the optimization objectives. Ant colony optimization algorthms are resorted to solving the MOCLO model and can approximate every part of Pareto-optimal front uniformly. A case study show that MOCLO model can give better cycle length than F-B method and ARRB method on the trade-off among various objectives and the alternative solutions for all potential situations.
出处
《同济大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2009年第6期761-765,共5页
Journal of Tongji University:Natural Science
基金
国家自然科学基金资助项目(70631002)
关键词
信号控制
周期时长
多目标优化
蚁群优化算法
signal control
cycle length
multi-objective optimization
ant colony optimization algorithms
