An improved discrete bees algorithm for correlation-aware service aggregation optimization in cloud manufacturing

As an effective approach to realize value and efficiency-added manufacturing activities, service aggregation usually plays an important role in cloud manufacturing (CMfg), and in order to improve the performance of manufacturing service aggregation, the quality of service (QoS) issue should be considered. However, most existing works related to QoS-based service aggregation assume that the services which to be aggregated are independent from each other, and the service aggregation evaluation models ignore the correlation between the services, which directly leading to inaccuracies in the evaluation of QoS of aggregation services. In this paper, three kinds of correlation are considered and the correlation-aware QoS model of aggregation service is presented in three levels. Furthermore, how to select an appropriate service to compose newly and optimal performance service from massive cloud services is discussed, as the service aggregation optimal selection problem is one of the key issues in CMfg. An improved discrete bees algorithm based on Pareto (IDBA-Pareto) is proposed to solve the problem in this context for CMfg. The presented method adopts a novel neighborhood searching mechanism underpinned by variable neighborhood searching (VNS) to improve the exploitation ability. The dynamic crowding distance adjustment strategy and the Pareto solution acceptance strategy at a certain probability are utilized to maintain diversity of solutions in population, so as to facilitate escaping from local optimum. The simulation results validate the effectiveness, high-efficiency, and superiority of IDBA-Pareto due to better population diversity and convergence speed.