Interval Type-2 Fuzzy Deep Reinforcement Learning-Based Operational Optimization of Industrial Aerodynamic System

Industrial aerodynamic system (IAS) provides the high-pressure air for consumer users, which is the primary source of power in the industrial park. Thus, their accurate measurement balance state and optimal scheduling are of paramount significance for reducing operational costs and enhancing operational efficiency. In order to solve this problem, an approach known as interval type-2 (IT2) fuzzy deep reinforcement learning (IT2FDRL) is proposed in this study, which is capable of integrating expert knowledge and operational data to enhance the performance of optimal scheduling in complex industrial scenarios. Moreover, the IT2 fuzzy line graph (IT2FLG) model is proposed to realize IAS balance state measurement which establishes a source-network-load-storage model by integration of physical and data-driven approach. To verify the effectiveness of the proposed approach, the performance of IT2FDRL is validated on the advanced process simulator (APROS) simulation platform with a typical operational scenario, the proposed method has salient advantages over the traditional methods in aspects of both operational cost and efficiency.