Optimization of Mechanical Manufacturing Processes Via Deep Reinforcement Learning-Based Scheduling Models

Abstract

With the development of Industry 4.0, intelligent manufacturing has become a prominent trend. This study focuses on applying artificial intelligence algorithms to optimize mechanical manufacturing processes, aiming to improve productivity, product quality, and reduce resource waste. We introduce an intelligent scheduling algorithm based on deep reinforcement learning for machinery manufacturing processes. The model utilizes deep Q-network (DQN) to make efficient production scheduling decisions and can handle complex and dynamic production environments. The experimental results demonstrate the algorithm's superior performance in both single-production line and multi-production line collaborative operations. Specifically, it achieves significant improvements in key performance metrics, such as production cycle time, resource utilization, order delay rate, and emergency order response time. Additionally, the algorithm showcases strong adaptability, effectively managing different types of orders and production lots. Quantitative improvements are observed in production cycle time and order delay rate, which highlight the practical benefits of the proposed approach in real-world applications.