基于FAHP与规划图融合的Web服务组合方法

近年来,随着云计算的发展,越来越多的服务被发布在网上。如何将不同的Web服务组合在一起并使其满足功能性需求和非功能性需求成为了一个研究难点。Web服务质量(Quality of Service,QoS)感知的Web服务组合问题属于NP难问题。为了解决这个问题,文中提出一种融合FAHP与改进Graphp lan算法的方法(FAHP and Improved Graphplan,FIGP)。首先,根据用户偏好使用模糊分析层生成服务的综合QoS;其次,在Graphplan向前扩展中,使用动态阈值对竞争力较差的服务进行剪枝,在保留关键服务的同时降低了时间复杂度;最后,在Graphplan向后搜索阶段,在满足功能性需求的前提下选择综合QoS最好的服务加入到组合中。实例分析和实验结果表明,与普通的Graphplan,Skyline及其他方法相比,FIGP不仅较好地提高了服务组合的质量,而且显著缩短了程序的执行时间。

Web Service Composition by Combining FAHP and Graphplan

In recent years,with the advance of cloud computing,more and more services have been published online.How to search an optimal composition with both functional and non-functional requirements has become a challenging problem.QoS-aware web service composition is an NP-hard problem.To solve this problem,a system combining FAHP with improved Graphplan algorithm was proposed.Firstly,the overall QoS of service is generated by using FAHP according to user preferences.Se-condly,in the forward expand stage of Graphplan,dynamic threshold is used to prune less competitive services,which reduces time complexity while ensuring that critical services are retained.Finally,in the backward searching stage of Graphplan,service with best overall QoS is selected into the composition,under the premise of meeting the functional requirements.Experimental results show that the proposed algorithm not only improves the quality of service composition,but also reduces the program running time significantly compared with the ordinary Graphplan,Skyline and other methods.