Abstract: In the process of service recommendation of Cloud Platform for cold start-up users,the cross-attribute of service resource hierarchy is not fully considered,which leads to the decision-making behavior of expert diversification and differentiation,furthermore,it is impossible to provide reasonable service ranking for cold-start users,so a service decision-making recommendation method based on language Z-number and cloud model was proposed.The attribute system of multi-party decision-making standard was constructed,which included platform users,cloud platform and service providers.Language Z-number was transformed into classical fuzzy number to determine the decision-making expert weight and the decision-making standard attribute weight.Z composition cloud model was used to get service decision recommendation list under language Z-number environment,which provided decision recommendation service for cold start users.Taking the decision-making recommendation ranking of the five service resources of “industrial manufacturing”,“software and informationization”,“environmental protection equipment”,“intelligent life equipment” and “special equipment” in the platform of “orange-cloud industrial product collaborative R & D” as an example,the feasibility and effectiveness of the proposed method were verified,and the theoretical basis of high-quality service was provided for cold-start users of cloud platform.

Key words: cloud platform, cold start user, multi-attribute decision making, language Z-number, cloud model

摘要： 针对面向冷启动用户的云平台服务决策推荐过程中,未充分考虑服务资源层次属性的交叉特性而导致的专家多元化和差异化的决策行为,进而无法准确为冷启动用户提供合理服务排序的问题,提出基于语言型Z-number和云模型的云平台服务决策推荐方法。首先,构建包括平台用户、云平台和服务提供方的多方决策标准属性体系;其次,转换语言型Z-number为经典模糊数,确定决策专家权重与决策标准属性权重;再次,利用Z合成云模型在语言型Z-number环境下得到服务决策推荐列表,为冷启动用户提供决策推荐服务;最后,以“橙色·云工业产品协同研发”平台中“工业制造”、“软件与信息化”、“环保装备”、“智能生活装备”以及“特种装备”五大服务资源的决策推荐排序为例,验证了所提方法的可行性和有效性,为云平台的冷启动用户提供高质量的服务理论基础。

关键词: 云平台, 冷启动用户, 多属性决策, 语言型Z-number, 云模型