Smart Manufacturing Scheduling System: DQN based on Cooperative Edge Computing

Abstract

In this paper, Deep Q-Network (DQN) was adopted to solve the Job shop Scheduling Problem (JSP) in the smart factory process. On the other hand, cloud computing has sensitive issues in the manufacturing process such as communication delay time and security problems. Research on various aspects of introducing an edge computing system to replace it has been conducted. We propose cooperative scheduling among edge devices in a Multi access Edge Computing (MEC) structure for scheduling without the help of a cloud center in a smart factory edge computing environment. Moreover, efficient DQN is used for experiments based on transfer learning data, and the proposed framework is compared and analyzed with existing frameworks from the perspective of provider a smart factory service.