Self-adaptive bat algorithm for large scale cloud manufacturing service composition

Abstract

In order to cope with the current economic situation and the trend of global manufacturing, Cloud Manufacturing Mode (CMM) is proposed as a new manufacturing model recently. Massive manufacturing capabilities and resources are provided as manufacturing services in CMM. How to select the appropriate services optimally to complete the manufacturing task is the Manufacturing Service Composition (MSC) problem, which is a key factor in the CMM. Since MSC problem is NP hard, solving large scale MSC problems using traditional methods may be highly unsatisfactory. To overcome this shortcoming, this paper investigates the MSC problem firstly. Then, a Self-Adaptive Bat Algorithm (SABA) is proposed to tackle the MSC problem. In SABA, three different behaviors based on a self-adaptive learning framework, two novel resetting mechanisms including Local and Global resetting are designed respectively to improve the exploration and exploitation abilities of the algorithm for various MSC problems. Finally, the performance of the different flying behaviors and resetting mechanisms of SABA are investigated. The statistical analyses of the experimental results show that the proposed algorithm significantly outperforms PSO, DE and GL25.