Abstract: There are a large number of cloud services on the cloud manufacturing platform. How to select satisfactory cloud services from a large number of cloud services for users without professional knowledge has become an important challenge. In the process of cloud service selection, when users are at different stages of the product life cycle, their preference for the attributes of cloud services is different. In other words, the preferences of users change dynamically over time, which will directly affect the results of cloud service selection. A cloud service selection method is proposed that considers the dynamic preferences of users. First, the intuitionistic fuzzy numbers is used to process heterogeneous QoS attribute data. Secondly, under the intuitionistic fuzzy environment, the entropy method is used to obtain the QoS attribute weight as the user's initial preference, and then the user's dynamic preference is determined by using the Markov chain. Finally, the VIKOR method is used to evaluate cloud services under an intuitionistic fuzzy environment, and a ranking is given. Experimental results show that this method can effectively help decision makers in different life cycles of products to choose satisfactory services, and the method proposed is reliable in different scenarios.

Key words: cloud service, intuitionistic fuzzy set, dynamic preference, Markov chain, VIKOR (vlsekriterijumska optimizacija I kompromisno resenje)