ADVANCED LINE CONTROL STRATEGIES: AN APPROACH WITH DEEP REINFORCEMENT LEARNING

This study investigates the application of Deep Reinforcement Learning (DRL) in optimizing production control in a multinational company, exploring the intersection between innovation and technology in Industry 4.0. By replacing the current production line logic with a Reinforcement Learning (RL) model, significant improvements were observed, especially in atypical scenarios. The DQL model demonstrated less production reduction compared to the current logic in simulations involving critical machine shutdowns. These results highlight the capability of DRL to handle complex and unpredictable challenges, standing out as a valuable tool for improving efficiency and competitiveness in the industry. Further exploration of reinforcement learning algorithms and neural network configurations is recommended for additional advancements in this field.