同时送取货的随机旅行时间车辆路径问题方法

Method for the stochastic traveling time VRPSPD problem

建立了同时送取货的随机旅行时间车辆路径问题(STT-VRPSPD)的机会约束规划模型,构建了分散搜索算法求解策略.分散搜索算法中,针对STT-VRPSPD问题的复杂特性,构造了解的改进策略、组合策略,并采用改进的节约算法构造分散搜索算法初始解,从而使文中设计的分散搜索算法更加适应STT-VRPSPD问题特有的负载波动性.仿真实验中,首先对分散搜索算法的参数设置进行分析,确定了最优参数组合;然后基于经典的Dethloff算例数据,构造了STT-VRPSPD的测试算例,并对分散搜索算法和遗传算法进行了对比分析,结果表明,分散搜索算法对于STT-VRPSPD的求解质量优于遗传算法.

This paper set up a chance-constrained programming model for STT-VRPSDP（stochastic traveling time vehicle routing problem with simultaneous pick-up and delivery）,designed respectively a scatter search algorithm applicable to this problem.In response to the complexity of STT-VRPSDP, this paper constructed innovatively an improvement method and a combination method derived from the basic theory of scatter search algorithm.Moreover,with C-W algorithm as an approach to get the initial solution to scatter search algorithm,the scatter search algorithm designed in this paper catered better to the loading floating feature particular to STT-VRPSDP and therefore could better the quality of solution. The simulation experiment firstly analyzed the parameters setting of scatter search algorithm,after which several sets of instances were chosen for the purpose of comparing and analyzing scatter search algorithm and genetic algorithm.The computation results show that solutions of scatter search algorithm are better than those of genetic algorithm.