VRP问题的研究起步较早,求解方法也非常丰富,然而,面对客户规模庞大,交通网络复杂的多约束车辆优化调度问题,现有算法显得无能为力.为有效解决需求点规模庞大的城市配送车辆优化调度问题,提出一种新的两阶段启发式算法——集束式算法,采用“集中后分派,分派后扩展”的思想,对末梢客户和同路段客户进行客户点合并,从全局上降低搜索范围,并提出相关客户点归并算法. However, in order to solve the problem of multi-constrained vehicle optimal scheduling with large-scale customers and complicated transport network, the existing algorithms seem powerless.In order to effectively solve the problem of large-scale urban delivery vehicles Optimization scheduling problem, this paper proposes a new two-stage heuristic algorithm called clustering algorithm, which adopts the idea of ​​“centralized allocation, extension after distribution” Reduce the search range, and put forward the relevant customer points and algorithm.