交叉口车流多目标运行效率决策研究  

Study of Multi-objective Operation Efficiency for Traffic Flow of Signalized Intersection

作　　者：陈娟[1] 耿俊杰[1] 贾明星[1] 

机构地区：[1]上海大学悉尼工商学院,上海201899

出　　处：《计算机仿真》2016年第4期200-205,共6页Computer Simulation

基　　金：国家自然科学基金资助项目(61104166);上海市教委科研创新项目资助(12YZ029)

摘　　要：研究交叉口多目标信号优化中的配时问题,由于客观环境的复杂性、系统存在模糊性等因素的影响,很难获得确定性决策信息来选择最优方案,提升车流运行效率。为解决上述问题,首先采用多目标超体积评价算法求解多目标信号配时模型,获得配时方案集。其次针对不确定型区间多属性决策问题,提出主客观信息偏差最小法,通过建立优化模型来处理主客观不确定性需求,确定配时方案。仿真结果表明,主客观信息偏差最小法能较好地权衡决策需求与决策矩阵客观信息,提高解的质量;根据多目标超体积评价算法和主客观信息偏差的优化与决策集成分析,综合考虑了不同车流运行评价指标,提升了交叉口整体运行效益。Aiming at solving the multi -objective signal timing problem for isolated intersections, this article presents a multi -objective optimization and decision making integration method （MODMIM） by taking the uncertain information into consideration. Firstly, the Hypervolume Estimation Algorithm for Multi -objective Optimization （HypE） is selected to solve the multi- objective optimization problem and obtain the Pareto front. Secondly, an in- terval multi -attribute decision making method called minimum deviation analysis for subjective and objective infor- mation （MDASOI） is proposed to select the satisfactory solution from the Pareto front. Simulation results show that the MDASOI can balance preference demand of decision - makers and objective evaluation information efficiently, and the MODMIM based on HypE and MDASOI can improve comprehensive traffic efficiencies effectively

关 键 词：多目标优化 多属性决策 区间 不确定性 信号配时 

分 类 号：TP237.4[自动化与计算机技术—检测技术与自动化装置]