Abstract: To solve the public problem namely cloud service providers'selection problem caused by combination of cloud computing technology and different industries such as medical,financial,energy and manufacturing,an index system was constructed by combining the views of technology and technological management.On this basis,an uncertain multi-attribute model of selecting providers was designed under the evidential reasoning framework.A determination method of deriving fair attribute weights from four dispersion measures of variance,deviation,entropy and CRiteria Importance Through Intercriteria Correlation (CRITIC) was developed by using this model,and the characteristics of the method were discussed and proved theoretically.The corresponding selection result of cloud service providers was further generated.The effectiveness and applicability of the proposed method was verified by a case.

Key words: cloud service, provider selection, objective weights, fair weights, evidential reasoning framework

摘要： 为解决医疗、金融、能源、制造等不同产业与云计算技术结合所面临的公共问题,即云服务供应商选择问题,从技术和技术管理两个视角构建了云服务供应商选择的指标体系。设计了证据推理框架下的不确定多属性供应商选择模型,依托该模型提出涵盖均方差、离差、熵、相关性四种离散度测度的公平属性权重确定方法,讨论并理论证明了方法的性质,进而产生相应的云服务供应商选择结果。通过实例验证了该方法的有效性与应用性。

关键词: 云服务, 供应商选择, 客观权重, 公平权重, 证据推理框架