A restructuring service cluster algorithm ABC optimised based on virtual resource selection probability

Virtual resources VR clustering represents a valuable advantage in terms of organisation, readability and efficiency for resource exploration. Service clusters often embed VR based on predetermined clustering methods e.g. hierarchical methods, partionning methods, fuzzy clustering, and density-based algorithms. Meanwhile, after several resource explorations and selections, the agent-based system, denoted cloud platform in cloud manufacturing CMfg, is able to gather statistical data concerning service requirements along with the probabilities of service clusters and their related VR to be selected. Consequently, the structure and definition of service clusters can be re-considered according to the probability of a set of VR to be selected. This article proposes a clustering method based on selection probability and ABC Artificial Bee Colony algorithm optimised. The present research also focuses on the global definition and improvements of the CMfg clustering process and the restructuring trigger function. Finally, the relevance and efficiency of the restructuring service cluster method are demonstrated in the precision and performance evaluation section.