刘必晶. 基于深度强化学习的综合能源系统优化调度[J]. 现代电力. DOI: 10.19725/j.cnki.1007-2322.2022.0296
引用本文: 刘必晶. 基于深度强化学习的综合能源系统优化调度[J]. 现代电力. DOI: 10.19725/j.cnki.1007-2322.2022.0296
LIU Bijing. Optimal Dispatch of Integrated Energy System Based on Deep Reinforcement Learning[J]. Modern Electric Power. DOI: 10.19725/j.cnki.1007-2322.2022.0296
Citation: LIU Bijing. Optimal Dispatch of Integrated Energy System Based on Deep Reinforcement Learning[J]. Modern Electric Power. DOI: 10.19725/j.cnki.1007-2322.2022.0296

基于深度强化学习的综合能源系统优化调度

Optimal Dispatch of Integrated Energy System Based on Deep Reinforcement Learning

  • 摘要: 针对综合能源系统中可再生能源和负荷的不确定性,提出一种基于深度强化学习的优化调度方法。首先,阐述了深度强化学习方法的基本原理;然后,提出了基于深度强化学习的综合能源系统优化调度模型,并对其中的状态空间、动作空间和奖励函数进行设计;继而,设计了基于异步优势策略梯度算法的模型求解流程;最后,通过算例仿真验证表明,所提方法能自适应源、荷不确定性,达到与传统数学规划方法相近的优化效果。

     

    Abstract: In allusion to the uncertainty of renewable energy and load in integrated energy system, an optimal dispatch method based on deep reinforcement learning was proposed. Firstly, the methodology of the deep reinforcement learning was expounded, and an optimal dispatch model based on the deep reinforcement learning, in which the state space, action space and reward function were designed, was proposed. Secondly, the model solving process based on asynchronous advantage actor-critic (abbr. A3C) algorithm was designed. Finally, the results of example simulation show that the proposed method can adaptively respond to the uncertainty of source and loads, and its optimization effect is similar to that of traditional mathematical programming method.

     

/

返回文章
返回