Game-Based Collaborative Scheduling With Fuzzy Uncertain Migration in Cloud Manufacturing

Cloud manufacturing (CMfg) provides on-demand services offered by the cloud platform to satisfy the individual requirements of users. However, the dynamics and uncertainty of the CMfg environment pose significant challenges to implementation of efficient synchronous scheduling of processing and logistic services. This paper proposes a game theory-based collaborative scheduling approach for effective utilization of distributed manufacturing and logistic resources with fuzzy uncertain task migration in CMfg, and the Nash equilibrium in this game theory is realized by a decision tree optimization algorithm. With this model, manufacturing and logistic services can cope with unexpected events whether or not task migration occurs. Moreover, the proposed approach takes into account independent and shared logistics, as well as delayed logistics, to improve the efficiency of transportation along the same route. Simulation results demonstrate that this approach is not only effective for the relevant optimization objective but also can achieve great performance under dynamic CMfg environments. Note to Practitioners—To formulate optimal production planning, how to solve the optimization of manufacturing and logistic resources is the main focus of smart manufacturing systems. In this paper, game theory is introduced to address the collaborative service scheduling issues in cloud manufacturing. Previous dynamic scheduling methods cannot further distinguish between independent logistics and shared logistics. Also, they rarely evaluate the impact of the task migration strategy on the occurrence of unexpected events from game point of view. Therefore, in this paper, a unique two-layer scheduling method based on game theory model that is composed of a processing service scheduling sub-game and a logistic service scheduling sub-game is presented. This model is implemented by adopting a proposed decision tree optimization algorithm that settles constrained non-linear programming. Simulation experiments show that the proposed method outperforms existing scheduling methods in terms of operational efficiency and fuzzy task migration decisions. Additionally, this method can be readily implemented and incorporated into real production settings. Future work could improve the proposed method by analyzing the uncertainties in point-to-point and hub-and-spoke networks across different transportation routes.