Abstract: In the cloud manufacturing environment, customers often face a huge amount of information and the lack of trust when they select cloud service resources. In view of the problem that how to select the trusted cloud service, a new method based on D-S theory is proposed. Six cloud service criteria are selected as the trusted characteristics set. The evaluation grade membership of each criterion is solved through the obtained qualitative and quantitative evaluation information, and then the utility preference degree of each criterion is derived as the basic probability assignment value of cloud service. The criteria weights are introduced to modify the basic probability assignment to reduce the conflict in the process of evidence synthesis. D-S theory synthesis rules are used to synthetize the evidences for trusted cloud services. The confidence interval of each cloud service is calculated to get the contrasting partial ordering relation of the cloud services. Finally, taking an auto parts manufacturer as an example, the feasibility and effectiveness of the proposed method are verified.

Key words: cloud service, D-S theory, credible evaluation, service selection

摘要： 云制造环境下，用户选择云服务资源时往往陷入信息量巨大以及信任缺失的困境。针对如何选取可信云服务的问题，提出了一种基于D-S理论的云服务评价方法。选取六个云服务指标作为可信特征集；通过获取定性定量指标评价信息，求解指标的评价等级隶属度；进而获得评价指标效用偏好度作为基本概率分配值；引入各云服务评价指标的权重修正基本概率分配以减少证据合成过程中的冲突；应用D-S理论合成规则对可信云服务的证据进行合成；计算各云服务的信任区间并将其进行偏序关系的对比，实现了云服务的选择。以某汽车零部件制造商选择云服务为例，验证所提方法的可行性和有效性。

关键词: 云服务, D-S理论, 可信评价, 服务选择