基于语义和动态QOS模型的Web服务匹配研究
发布时间：2018-07-31 07:38
【摘要】： 由于Web服务的种种优越性及广阔的应用前景,自诞生之日起就受到许多机构和公司的认可与推崇,许多组织机构已将Web服务引入他们的商业流程中,并获得了直接的经济效益。 然而随着Web服务的发展,网络上可用的Web服务会越来越多,如何从数量庞大的服务群中为用户找到满足其所需功能的Web服务,即Web服务的匹配已经成为一个亟待解决的问题。目前UDDI的Web服务发现机制还是基于关键字的匹配,从而导致Web服务发现的低精确性,同时,查询关键字可能与注册的Web服务的描述在语义上相同,而语法上不同,从而导致了低查全率;另一方面,随着Web服务的大量涌现,功能相似的服务会越来越多,而这些服务的非功能属性,如服务收费,服务的可靠性、稳定性等方面可能存在很大的差异;传统的匹配机制缺乏Web服务质量支持以实现最佳服务的发现。另外,用户对这些非功能性能的关心程度可能也有差异,如何按照用户的个性化要求为其找到认为性能最优的服务也是服务匹配中应该考虑的重要问题。 针对传统匹配机制的弊端,本文将本体技术引入到Web服务中,设计了一个基于语义的Web服务匹配模型,提出一套多层次、细粒度的匹配策略,改进了语义相似度的计算方法,具体工作包括: 第一,研究如何定义和表示Web服务。构建Web服务语义描述模型,使其支持对服务质量描述,通过把服务语义描述和服务质量描述以tModel的形式注册到UDDI中心,实现服务语义描述在UDDI中心的使用。 第二,传统的Web服务匹配模型已不能适应复杂的语义匹配的需要,本文提出基于语义和动态QOs的Web服务匹配模型,并进行了详细的层次结构设计,将模型分为四个层次和服务发布模块、服务查询模块、OWL-S/UDDI映射等若干子模块,对每个模块都进行了功能设计。 第三,针对如何从众多的功能相似的Web服务中根据服务质量发现最佳服务的问题,提出了新的服务质量描述模型,并且在综合考虑web服务质量的历史统计信息的基础上,提出了一个有效的Web服务质量评价模型并对服务质量综合指标的评价算法进行了详细的设计。为后续的服务描述研究和服务匹配研究提供基础。 第四,根据智能化和高效性的要求,研究怎样基于语义描述和本体论对广告服务描述与请求服务描述进行匹配。提出了一种多层次、细粒度的服务匹配策略,通过服务类别、服务基本描述信息(服务名称、服务描述)、服务I/O描述信息、服务质量四级匹配,返回满足用户请求的服务;并对词语相似度和语义相似度的算法进行了改进。 最后,本文对提出的模型进行了模拟实现,对改进的语义相似度等算法进行有效性验证和性能测试,结果表明,本文提出的模型能够很好地支持web服务的语义匹配,改进的算法一定程度上提高了服务匹配的准确性和查全率。
[Abstract]:Because of its advantages and broad application prospects, Web services have been recognized and respected by many organizations and companies since its birth. Many organizations have introduced Web services into their business processes and obtained direct economic benefits. However, with the development of Web services, there will be more and more Web services available on the network. How to find Web services to meet the needs of users from a large number of service groups, that is, the matching of Web services, has become an urgent problem to be solved. At present, UDDI's Web service discovery mechanism is still based on keyword matching, which leads to the low precision of Web service discovery. At the same time, query keywords may be semantically the same as registered Web services, but syntax is different. This leads to low recall rates; on the other hand, with the proliferation of Web services, there are more and more similar services, and the non-functional attributes of these services, such as service charges, service reliability, There may be great differences in terms of stability, and the traditional matching mechanism lacks Web QoS support to realize the discovery of the best service. In addition, the degree of concern of users to these non-functional performance may also be different. How to find the best performance services for users according to their personalized requirements is also an important issue that should be considered in service matching. Aiming at the disadvantages of traditional matching mechanism, this paper introduces ontology technology into Web services, designs a Web service matching model based on semantics, proposes a set of multi-level, fine-grained matching strategies, and improves the method of calculating semantic similarity. The specific work includes: first, how to define and represent Web services. The Web service semantic description model is constructed to support the service quality description. The service semantic description and the service quality description are registered to the UDDI center in the form of tModel to realize the use of the service semantic description in the UDDI center. Secondly, the traditional Web service matching model can no longer meet the needs of complex semantic matching. In this paper, a Web service matching model based on semantic and dynamic QOs is proposed, and the hierarchical structure is designed in detail. The model is divided into four levels and service release module, service query module, OWL-Sr / Uddi mapping and other sub-modules, each module is designed. Thirdly, aiming at the problem of how to find the best service according to the quality of service from many similar Web services, this paper proposes a new service quality description model, and on the basis of synthetically considering the historical statistical information of web service quality, In this paper, an effective Web service quality evaluation model is proposed and the algorithm for evaluating the comprehensive quality of service index is designed in detail. It provides the foundation for service description research and service matching research. Fourthly, according to the requirement of intelligence and efficiency, this paper studies how to match the description of advertising service with the description of request service based on semantic description and ontology. A multi-level, fine-grained service matching strategy is proposed, which returns the service satisfying the user's request through service category, service basic description information (service name, service description), service I / O description information, service quality four-level matching. The algorithm of word similarity and semantic similarity is improved. Finally, the proposed model is simulated and implemented, and the improved semantic similarity algorithms are validated and tested. The results show that the proposed model can support the semantic matching of web services well. The improved algorithm improves the accuracy and recall of service matching to some extent.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2008
【分类号】：TP393.09
本文编号：2154884
[Abstract]:Because of its advantages and broad application prospects, Web services have been recognized and respected by many organizations and companies since its birth. Many organizations have introduced Web services into their business processes and obtained direct economic benefits. However, with the development of Web services, there will be more and more Web services available on the network. How to find Web services to meet the needs of users from a large number of service groups, that is, the matching of Web services, has become an urgent problem to be solved. At present, UDDI's Web service discovery mechanism is still based on keyword matching, which leads to the low precision of Web service discovery. At the same time, query keywords may be semantically the same as registered Web services, but syntax is different. This leads to low recall rates; on the other hand, with the proliferation of Web services, there are more and more similar services, and the non-functional attributes of these services, such as service charges, service reliability, There may be great differences in terms of stability, and the traditional matching mechanism lacks Web QoS support to realize the discovery of the best service. In addition, the degree of concern of users to these non-functional performance may also be different. How to find the best performance services for users according to their personalized requirements is also an important issue that should be considered in service matching. Aiming at the disadvantages of traditional matching mechanism, this paper introduces ontology technology into Web services, designs a Web service matching model based on semantics, proposes a set of multi-level, fine-grained matching strategies, and improves the method of calculating semantic similarity. The specific work includes: first, how to define and represent Web services. The Web service semantic description model is constructed to support the service quality description. The service semantic description and the service quality description are registered to the UDDI center in the form of tModel to realize the use of the service semantic description in the UDDI center. Secondly, the traditional Web service matching model can no longer meet the needs of complex semantic matching. In this paper, a Web service matching model based on semantic and dynamic QOs is proposed, and the hierarchical structure is designed in detail. The model is divided into four levels and service release module, service query module, OWL-Sr / Uddi mapping and other sub-modules, each module is designed. Thirdly, aiming at the problem of how to find the best service according to the quality of service from many similar Web services, this paper proposes a new service quality description model, and on the basis of synthetically considering the historical statistical information of web service quality, In this paper, an effective Web service quality evaluation model is proposed and the algorithm for evaluating the comprehensive quality of service index is designed in detail. It provides the foundation for service description research and service matching research. Fourthly, according to the requirement of intelligence and efficiency, this paper studies how to match the description of advertising service with the description of request service based on semantic description and ontology. A multi-level, fine-grained service matching strategy is proposed, which returns the service satisfying the user's request through service category, service basic description information (service name, service description), service I / O description information, service quality four-level matching. The algorithm of word similarity and semantic similarity is improved. Finally, the proposed model is simulated and implemented, and the improved semantic similarity algorithms are validated and tested. The results show that the proposed model can support the semantic matching of web services well. The improved algorithm improves the accuracy and recall of service matching to some extent.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2008
【分类号】：TP393.09
【引证文献】
相关硕士学位论文 前1条
1 卫格格;一种支持QoS约束的Web服务组合的研究[D];西北大学;2010年
，本文编号：2154884
本文链接：https://www.wllwen.com/wenyilunwen/guanggaoshejilunwen/2154884.html
