Research on Intelligent Cloud Manufacturing Resource Adaptation Methodology Based on Reinforcement Learning

Abstract

In cloud manufacturing, an allocation platform integrates resources and assign tasks manufacturers. Improper allocation algorithm causes wasting of resources and budget. In this paper, a new bilateral adaptation algorithm based on Q-learning and improved Gale-Shapley algorithm is proposed for the manufacturer-dealer bilateral adaptation problem in the intelligent cloud manufacturing environment. The main idea of the algorithm is to first construct a manufacturer-to-dealer distance-based ordering model as well as a dealer-to-manufacturer Quality of Service based (QoS-based) ordering model, and then, a basic bilateral allocation result is obtained by the improved Gale-Shapley algorithm, which will allocate dealers and manufacturers into pairs depend on the overall cost and degree of satisfaction. At last, with the obtained pairing scheme, a more rational pairing approach can be obtained through the self-learning process in the Q-learning algorithm. Numerical experiment has been done and the proposed algorithm are compared with some traditional approaches. The experimental results show that the improved Gale-Shapley algorithm proposed in this paper obtains better results than other ways.