Web服务的语义关系挖掘与组合方法的研究

发布时间：2019-01-05 04:02

【摘要】：随着大数据时代的到来，Web服务技术得以进一步发展。稳定易用的Web服务大量涌现，如何将单一的、功能有限的Web服务组合起来为用户提供综合、满足个性化需求的组合服务成为了当前的研究热点。本文以服务类为对象，服务关系为支撑，通过挖掘服务类之间的语义关系来实现Web服务的自动服务组合。 1、研究Web服务的语义关系挖掘。定义Web服务之间的语义关系，通过服务聚类完成服务库的服务类划分。以Web服务的QoS属性为筛选标准，通过原子服务筛选以及服务QoS维度偏好的索引建立完成服务类的优化。以服务类为基础，提升服务关系挖掘的粒度，提出服务类的语义关系挖掘算法，，为自动服务组合方法提供语义关系支持。 2、研究Web服务的自动服务组合方法。自上而下将服务组合过程分为服务类的组合以及服务的绑定两个阶段。服务类的组合阶段使用规划算法完成语义关系图的构建，将服务组合映射为最短路径问题来求解服务类的组合方案。在组合方案的服务类节点上提出个性化的服务选择方法完成服务的绑定。 3、研究Web服务的服务选择方法。为满足用户的个性化需求，提出基于偏好推荐和QoS的服务选择算法。通过计算服务请求者和历史评价用户之间的评价偏好相似度来获取推荐用户，计算推荐用户相应候选服务的服务推荐度，并结合服务请求者的QoS维度偏好计算候选服务的综合效用值，完成个性化的Web服务选择。通过实验方法，分别对服务类语义关系挖掘、服务组合和服务选择的算法进行验证，结果分析表明本文设计的算法是可行和有效的。研究成果能有效解决服务网络中原子服务及其相互之间的语义关系海量增长的问题，并提供个性化和优质的组合服务，进一步提高服务请求的用户满意度。
[Abstract]:With the arrival of big data, Web service technology has been further developed. A large number of stable and easy-to-use Web services have emerged. How to combine the single and limited Web services to provide comprehensive services for users to meet the needs of personalized composition has become a hot research topic. This paper takes the service class as the object, the service relation as the support, by mining the semantic relation between the service classes to realize the automatic service composition of the Web service. 1. The semantic relation mining of Web services is studied. The semantic relationship between Web services is defined and the service cluster is used to partition the service class of the service library. With the QoS attribute of Web service as the filter criterion, the optimization of service class is accomplished by atomic service filtering and the index of service QoS dimension preference. Based on the service class, the granularity of the service relationship mining is improved, and the semantic relation mining algorithm of the service class is proposed, which provides the semantic relationship support for the automatic service composition method. 2. The automatic service composition method of Web service is studied. The process of service composition is divided into two stages: service class composition and service binding. In the composition phase of the service class, a programming algorithm is used to complete the construction of the semantic relational graph, and the service composition is mapped to the shortest path problem to solve the composition scheme of the service class. A personalized service selection method is proposed on the service class node of the composition scheme to complete the service binding. 3. The service selection method of Web service is studied. A service selection algorithm based on preference recommendation and QoS is proposed to meet the individual needs of users. By calculating the similarity of evaluation preference between the service requester and the historical evaluation user, the recommended user is obtained, and the service recommendation degree of the corresponding candidate service of the recommended user is calculated. Combined with the QoS dimension preference of the service requester, the comprehensive utility value of candidate service is calculated to complete the personalized Web service selection. The algorithms of service class semantic relation mining, service composition and service selection are verified by the experimental method. The results show that the proposed algorithm is feasible and effective. The research results can effectively solve the problem of massive growth of atomic services and their semantic relationships in service networks, and provide personalized and high-quality composite services to further improve the user satisfaction of service requests.
【学位授予单位】：上海大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09;TP391.1

【参考文献】