Abstract: The logistics enterprises in our country is that the scale for a logistics enterprise is small, the market share small, the function single, each one does business alone. The strengths of different company have not been combined yet, it is not in conformity with the modern logistics’ task. In view of the above issues, BPEL workflow technology make the services provided by different companies which selected by every node combine a work flow. The problems in two aspects have been solved by using the improved particle swarm optimization: how to rapidly select a optimal combination of services in subset for service waited for selection of each node to meet the requirements; how to decide the front and rear order when a single workflow can’t service many customers. The results show that this combination can realize a reasonable combination of service provided by a different logistics company to improve the overall logistics service level.

Key words: modern logistics, workflow, improved particle swarm algorithm, combination of services, service ensemble waiting for selection

摘要： 我国的物流企业呈现了规模小，市场份额小，各自为战，功能单一的现象。各公司优势不能结合起来，不符合现代物流的要求。针对以上问题，基于bpel工作流技术，将每个物流节点选择的不同公司提供的服务组合成一个业务流程。用改进的粒子群算法解决了两个问题：如何在各个节点的服务候选集中快速选取一套符合要求的最优服务；在不能同时服务多个用户的情况下，按怎样顺序服务用户。实现了服务的合理组合，提高整体的物流服务水平。

关键词: 现代物流, 工作流, 改进的粒子群算法, 服务组合, 服务候选集