Scientific Development Research

This paper proposes a multi-time-scale dispatch strategy for integrated energy systems based on deep reinforcement learning, aiming to enhance the efficiency and reliability of energy system dispatch. By constructing a multi-time-scale model and integrating deep reinforcement learning techniques, the strategy enables dynamic optimization of dispatch decisions for energy production, transmission, and consumption. Simulation and experimental results demonstrate that the proposed strategy outperforms traditional methods when dealing with complex system dynamics, not only reducing system costs but also improving energy utilization and flexibility. The method shows strong potential for practical applications, particularly in the context of smart grids and renewable energy integration, enabling more refined dispatch optimization.

Deep reinforcement learning; Integrated energy system; Multi-time-scale; Dispatch optimization; Energy management