针对基于图搜索实现Web服务自动组合存在搜索空间太大以及难以表达各种组合结构的问题,提出一种基于语义匹配关系确定组合结构的方法。该方法首先对Web服务进行形式化的语义描述,然后基于语义匹配关系,将服务库中只与用户请求的输入输出关联的服务构成语义关系图。在此基础上,基于语义匹配关系定义各种组合结构模型,以综合语义匹配度为寻优目标改进广度优先搜索算法,定义不同结构的语义匹配度计算方法,生成一条综合语义匹配度最优的Web服务组合路径。通过实例验证了该算法的可行性。
The method based on graph search was a simple and direct way to realize the Web service automatic composition, hut the search space was too large and it was difficult to express various combination structures among serv ices. To solve this problem, a method based on semantic matching relationship to determine combination structure was presented. Formal description of the Web services semantics were presented, and then based on semantic matehing relationship, the semantic relationship graph was established by services which were only related to user provided input and expected output. On this basis, various combination structure models were defined based on semantic matching relationship, and taking integrated semantic matching degree as optimal goal, the breadth-first search algorithm was improved, the calculation method of semantic matching degree in various combination structure was defined, and the service combination path which owned the optimal integrated semantic matching degree was generated. The feasibility of proposed algorithm was verified though an example.