依赖知识的服务组合算法

A Knowledge-Based Service Composition Algorithm

针对面向服务的体系结构中Web服务和服务组合技术在下一代互联网和云计算领域中的应用问题,提出了一种依据知识对原子服务进行组合的3步服务组合算法,分为相邻节点连接匹配、断枝节点清理和原子服务组合3个步骤。相邻节点连接匹配就是完成相邻的服务节点之间连接匹配,将所有允许的连接都存入弧的连接链表中。通过断枝节点清理,删除所有没有连接的候选原子服务,因为在候选队列中会存在许多断枝服务节点。通过原子服务组合,从清理过断枝节点之后的剩余候选服务中整理出服务组合实例。实验表明:算法可以实现服务的自动化装配,不需要用户参与服务调用过程,通过原子服务间的可连接性以及知识模型生成服务组合实例,从而实现原子服务数量和检测范围的有效控制。

A knowledgebased service composition algorithm is proposed to apply the techniques of serviceoriented architecture including Web service and service composition in the area of the next generation Internet and the cloud computing. The algorithm consists of three key steps, i.e. connection matching between adjacent nodes, deleting broken nodes, and building service composition instance. All possible links between candidate primary service nodes are found and stored in an arc list at the first step. Then, the candidate primary service nodes that can not make a legal path are deleted in the second step. A service composition instance is built from the rest candidates in the step 3. The algorithm automatically assembles services without user’s intervention during the service procedure, and builds the service composition instance in terms of a knowledge model and the connectivity between primary services. Test results show that the number of primary services and detection ranges can be effectively controlled.