Abstract: In 4th party logistics platform, transportation, warehousing and other services were selected from massive services for service composition, so that to provide a solution for client's needs of goods dispatching. To improve the quality of logistical composite service and ensure the reliability and freshness of goods dispatching, a three stage optimization algorithm for dynamic service composition was proposed by aiming at logistic service quality. Subjective and objective factors were considered to obtain synthetic weights, and Quality of Service(QoS)evaluation model of logistical service reflecting users'preferences was introduced. According to established service selection strategies, the algorithm simulated the process of ants searching route to find out several better routes, and a network of services relevance was formed for services in these routes. By using greedy strategy, the most suitable logistics route was obtained. Performance evaluations showed that the algorithm could efficiently and dynamically obtain high quality logistics composite services which satisfy customer's requirements.

Key words: massive services, logistics, service quality, service composition, combinatorial optimization algorithm, search algorithms

摘要： 第四方物流平台中需要从海量服务中挑选运输、仓储等服务并组合在一起，为客户的货物调度需求提供一个解决方案。为提高组合服务的效率和质量，保证货物调度的可靠性和实效性，提出以物流服务质量为目标的三段式动态服务组合优化算法。首先综合主客观权重，构建了能体现用户偏好的物流服务质量评价模型。算法改进了蚁群算法搜索路径的思想，按照制定的服务筛选策略找出多条较优的物流路径（组合服务），然后建立这些路径中服务的关联网，在关联网上通过贪心选择搜索出最合适的物流路径。性能评价表明，所提算法能够较高效地动态组合出满足需求的高质量物流服务。

关键词: 海量服务, 物流, 服务质量, 服务组合, 组合优化算法, 搜索算法