基于灰色关联分析的Web服务选择

Web service selection based on grey correlation analysis

为方便用户选择最优Web服务,利用灰色系统理论对Web服务质量QoS属性因子进行分析,提出了一种基于用户QoS偏好的Web服务灰色关联分析方法。考虑到Web服务QoS的不确定性,该方法使用区间对Web服务QoS值进行建模。为了确定候选服务的QoS与用户QoS需求的符合程度,先针对服务的每个QoS属性,计算其与用户QoS需求的灰色区间关联系数;然后结合各个QoS属性的关联系数计算候选服务的QoS与用户QoS需求的综合灰色区间关联度,关联度越大的服务越符合用户的要求;最后从满足用户功能需求的Web服务中选择灰色关联度最大的Web服务推荐给用户。与其它Web服务评价模型相比较,该模型更加符合Web服务QoS的实际情况,能够在服务QoS信息不充分、不确定的环境下,对QoS属性进行合乎实际的分析处理,从而得到更加合理有效的QoS评价。

To help users select the service that best satisfies their non-functionality requirements from a set of Web services, we propose a grey correlation analysis based QoS-aware service selection method in this paper via analyzing service QoS attribute factors using the grey system theory. Based on that service QoS information is usually uncertain and incomplete, the proposed method uses intervals to model the QoS attribute values of Web services. To determine how well a service satisfies users＇ con- cerned QoS requirements, the method firstly adopts a set of functions to normalize interval grey numbers of services＇ QoS on various QoS attributes with different metrics and scales. Then it computes services＇ grey incidence degree coefficients of interval grey numbers on each QoS aspects. Finally it combines each service＇s grey incidence degrees on all QoS attributes to obtain an overall grey incidence degree. The service with the largest grey incidence degree is recommended to users. Compared with other Web serv- ice evaluation models, our approach is more suitable for real Web service systems where QoS information is uncertain and incomplete, and it can provide more effective and reasonable evaluation for Web service selection.