在对基本车辆调度问题(VRP)进行简单描述的基础上,提出了求解该问题的两类混合策略.这两类策略均采用两阶段的思想:第一阶段,利用节约法对顾客进行聚类;第二阶段,构造适合求解VRP的改进dynasearch算法,且将基于随机kick的迭代局域搜索算法与改进dynasearch算法相结合,提出了迭代dynasearch算法.通过实验仿真,验证了这两类策略在很大程度上能改进问题的解.
The basic vehicle routing problem (VRP) is considered in this paper. Two kinds of mixed methods are proposed for this problem. Both two methods adopt two-phase strategy. In the first phase, savings method is used to assign customers to vehicles. In the second phase, we construct improved dynasearch algorithm for VRP, and combine dynasearch with the iterated local search algorithm, which is named as iterated dynasearch algorithm. From the results of extensive computational tests, both two methods can improve the solution greatly compared to the traditional one.