To balance the stochastic power disturbance in integrated energy system (IES), a novel automatic generation control (AGC) dispatch is proposed by taking account of the regulation rule that applies to a performance-based frequency regulation market, with the aim to reduce area control deviation and regulation mileage payment while complying with constraints of various regulation units. Thus, a multiple experience pool replay twin delayed deep deterministic policy gradient (MEPR-TD3) is put forward to improve the training efficiency and the action quality via four improvements including the multiple experience pool probability replay strategy. Finally, the performance of the proposed algorithm is verified on an extended two-area load frequency control (LFC) model and Hainan province IES with different demand of multiple energy.

为了平衡集成能源系统（IES）中的随机功率干扰，提出了一种新颖的自动发电控制（AGC）调度，该调度考虑了适用于基于性能的频率调节市场的调节​​规则，旨在减少区域控制遵守各种法规单位的约束条件，并支付偏差和法规里程。因此，提出了一种多重体验池重播双延迟深度确定性策略梯度（MEPR-TD3），通过包括多重体验池概率重播策略在内的四项改进来提高训练效率和动作质量。最后，在扩展的两区域负荷频率控制（LFC）模型和具有多种能源需求的海南省IES上，验证了该算法的性能。