Optimizing Smart Business Operations Using Deep Q-Networks (DQN) and IoT Data

Abstract

Internet of Things (IoT) infrastructures in the form of smart businesses are used to monitor and optimize smart business operations, but the traditional operational management system does not easily cope with the dynamic changes of the industrial environment. This study aims to develop a smart business optimization framework using the Deep Q-Network (DQN) reinforcement learning algorithm that can be combined with manufacturing analytics using IoT to enhance manufacturing operational efficiency, resource utilization, predictive maintenance, and energy management. All the parameters of IoT such as temperature, vibration, energy consumption, latency, maintenance score, and production efficiency were simulated in the dynamic enterprise operational environments using the Intelligent Manufacturing Dataset from Kaggle. To realize adaptive operational decision-making, the data preprocessing, state space modeling, reward function optimization, and DQN policy learning method are proposed. The operational efficiency, optimization accuracy, convergence of the reward, reduction in downtime, and energy optimization metrics were used for experimental evaluation. The proposed framework was able to optimize the accuracy of the system with 97.68%, operational efficiency with 94.83%, and utilization of resources with 92.46%, whereas it reduced the downtime by 68.40% and energy consumption by 27.69%. Comparative analysis was also performed, showing that the proposed DQN framework had better performance than the traditional machine learning models such as Artificial Neural Network (ANN), Random Forest (RF), and Q-Learning strategy models. The research offers a smart and scalable reinforcement learning model that can be applied to the use of Industry 4.0 businesses by optimizing the autonomous operations of such businesses with enterprise intelligence modeled using IoT.