Deep Reinforcement Learning Based Acceleration Approach for Day-Ahead Optimal Dispatch of Integrated Energy Systems

As the energy revolution proceeds, integrated energy systems (IESs) are becoming increasingly indispensable. However, the economic dispatch problem of IESs is generally formulated as a complex mixed-integer nonlinear programming problem (MINLP) with various nonlinear constraints, which is difficult to solve. In this paper, we propose a deep reinforcement learning (DRL) based acceleration approach to deal with these nonlinear constraints. Thus, the original MINLP could be transformed into a mixed-integer linear programming problem (MILP) which can be tractably handled by existing optimization techniques. Numerical results have verified the effectiveness of the proposed strategy.