针对如何在海量Web服务中进行快速的服务自动组合、以满足客户复杂需求的问题,提出一种基于规划图思想的搜索组合算法,该算法在正向扩展阶段除了依据所选取服务的功能与非功能属性进行有效剪枝外,还实时级联更新已驱动服务的QoS信息并有效记录所操作服务的前驱服务,通过前驱索引集合,确保逆向求解阶段的效率并满足了QoS最优的约束。在实验环节证明了注册服务数量与求解效率的线性关系,以及与其他算法相比较所体现出来的性能优势。
For a quick service automatic composition in the vast Web service, to meet the demand of customer complex problems, it puts forward a graphplan based algorithm of the components search. The algorithm prunes effectively according to the function of the selected service and non-functional attributes in the forward extension stage. Besides, it real-time cascade updates enabled service QoS information and records the precursor service of operation service effectively, to ensure the efficiency of the reverse search phase and satisfy the QoS of the optimal constraint by precursor index collection.Experimental evaluation of linear relationship is made between registration service quantity and efficiency of solving, and the better performance advantages than other algorithms.