Web服务组合是服务计算领域的研究热点之一,面对当前因特网络中存在的大量不同服务类型的Web服务,现有的服务组合方法在服务自治协作方面仍存不足。为有效进行服务组合,提高组合性能和质量,提出一种服务Agent合作网络模型(CONETM)以及相关合作算法,使得一个Agent管理一个或多个相同服务类型的Web服务,并通过建立动态Agent合作关系实现对Web服务的最优组合。实验结果显示,CONETM方法比简单随机组合和分布式Agent联盟方法在组合时间、通信消息以及用户服务满足率等性能方面更好。
Web services composition is a hot research topic in services computing field. Confronting a mass of Web services owning different types of services in current Internet, existing services composition methods are yet insufficient in the capacity of autonomous collaboration. In order to efficiently compose the services for improving the performance and quality of composition, we propose a service Agent collaboration network model (CONETM) and the related collaborative algorithms, which let an Agent manages one or more Web services with same type of service and fulfil the optimal composition of Web services through forming dynamic collaborative relation between the Agents. Experimental results show that the proposed CONETM can get better performance in aspects of composition time, communication messages and users' satisfactory rate of services than the methods of simple random composition and distributed agent coalition.