A modified discrete invasive weed algorithm for optimal service composition in cloud manufacturing systems

Abstract

Cloud manufacturing (CMfg), as a new service-oriented manufacturing paradigm, is aiming towards sharing and collaborating among distributed manufacturing resources and capabilities. As a result, selecting and combining these services into a composite service to meet the user’s requirements while keeping up the optimal service performances is of paramount importance. In this paper, first of all, QoS-aware service composition and optimal selection (SCOS) problem is formulated as an optimization problem and then a modified discrete invasive weed algorithm is proposed and applied as a new approach for solving the NP-hard SCOS problem in CMfg context. The algorithm which mimics the process of weed colonization and distribution, is modified by putting into effect a novel mechanism for mapping normally-distributed dispersal values into the mutation probability of corresponding dispersal direction. The experimental results prove the good performance and robustness of the approach.