Abstract: Due to the optimization problem of dynamic vehicle routing problem, an improved algorithm for DVRP（Dynamic Vehicle Routing Problem） is proposed. Based on analyzing local features of routing optimization problem, the method uses variable neighbourhood search heuristics to explore, and combines mutation mechanism to exploit. The dynamic routing blocks are inserted into vehicles according to neighbor priority principle. Experimental results show that the improved algorithm is effective.

Key words: Dynamic Vehicle Routing Problem（DVRP）, variable neighbourhood search heuristics, mutation mechanism, neighbor priority principle

摘要： 针对动态车辆路径问题DVRP（Dynamic Vehicle Routing Problem）的优化问题，提出一种改进算法。该算法在分析路径寻优问题的局部特性的基础上，利用变邻域搜索算法VNS（Variable Neighbourhood Search）对路径空间进行“局部探索”，结合变异机制对路径空间进行“全局开采”，最后根据近邻优先原则将动态路径片段安插到适宜的路径中。实验结果验证了算法的有效性。

关键词: 动态车辆路径问题, 变邻域搜索算法, 变异机制, 近邻优先原则