基于两阶段求解算法的动态车辆调度问题研究  

Research on dynamic vehicle routing problem based on two-phase algorithm

作　　者：王旭[1] 葛显龙[2] 代应[3] 

机构地区：[1]重庆大学贸易与行政学院,重庆400044 [2]重庆大学机械工程学院,重庆400044 [3]重庆理工大学工商管理学院,重庆400050

出　　处：《控制与决策》2012年第2期175-181,共7页Control and Decision

基　　金：国家863计划项目(2006AA04A123);重庆市自然科学基金项目(CSTC.2008BB2173);教育部人文社会科学研究青年基金项目(09YJC630247)

摘　　要：在分析需求动态变化的基础上,根据需求信息的提出顺序,将动态配送问题转换成不同时刻的静态车辆调度问题,建立基于时间轴的动态车辆调度模型;利用量子理论改进遗传算法,设计量子遗传算法;针对动态车辆调度问题实时性强的特点,设计"初始优化阶段+实时优化阶段"的两阶段求解策略,通过信息更新插入动态需求客户,并对已产生的计划路径进行局部优化调整.通过仿真计算,验证了模型和算法的有效性.On the basis of analyzing dynamic needs information of customers,and according to the dynamic order information,the dynamic vehicle scheduling model based on the time taxes is established.The dynamic distribution problem is transformed into a series of static distribution problems.At the same time,quantum genetic algorithm is designed for vehicle routing problem（VRP）.Aiming at the real time of dynamic vehicle scheduling problem,the two-phase solution of ＂initial optimization stage＂ and ＂real-time optimization stage＂ are established,which optimizs sub-routes through continuously updating information and inserting the dynamic needs customers.Simulation results show the effectiveness of the model and algorithm.

关 键 词：量子遗传算法 动态车辆调度问题 两阶段算法 时间轴 

分 类 号：F274[经济管理—企业管理]