一种使用改进熵权TOPSIS的Web服务质量评价排序方法

Service Ranking Method Based on Improved Entropy TOPSIS

服务质量评价排序是服务计算领域重要的研究问题之一.当前研究中对于QoS各维属性赋权过程多使用专家系统或直接用户指定,这种方法存在主观性过强的问题,同时不能发现在实际中对服务整体表现有重大影响的QoS属性维度.针对此问题,本文提出了一种基于改进熵权TOPSIS的服务质量评价排序方法,采用了客观赋权的熵权法计算QoS各维属性对应的客观权重,并在此基础上与主观权重相结合,既提高了决策过程的客观性,又可以更好的贴近用户个性化需求.通过将服务排序转换为多目标决策问题,利用熵权改进传统TOPSIS方法可以得到服务质量优劣排序.最后通过一个排序实例证明该方法的可行性和有效性.

Web service ranking based on QoS constraints is an essential aspect of web service discovery, selection, composition and recommendation. Several models and algorithms for considering QoS requirement for service ranking have been proposed. However, these methods usually only decide the QoS attribute weights by expert system or directly using user＇s order,which may cause the too strong subjectivity problem. Aim this problem, we present in this paper a new web service ranking method based on improved Entropy- TOPSIS. The entropy part can dig out the real influence on the overall web service performance from different service attributes dimensions. And at the TOPSIS part, transfer the service ranking problem to multi-objective decision model, then the ordering queue of service quality can be gained by using entropy coefficient to improve the traditional TOPSIS. The efficacy of our method is validated by both mathematical analysis and simulation experiments.