摘要
设计了一种多目标的电网调度模型,在保障电网安全运行的基础上,实现最大化新能源消纳和最小化运行成本的目标。引入深度确定性策略梯度(deep deterministic policy gradient,DDPG)算法与环境交互得到最优调度策略。针对DDPG算法存在的训练不稳定的问题,提出一种规则引导DDPG算法,并在训练过程中加入双价值网络。实验结果表明,所提方法能够更好地实现调度目标,在原DDPG算法的基础上提高了模型稳定性和有效性。
A multi-objective power grid dispatching model was designed to maximize new energy consumption and minimize the operation cost on the basis of ensuring the safe operation of the grid.In addition,a deep deterministic policy gradient(DDPG)algorithm was introduced to obtain the optimal scheduling policy through interaction with the environment.Aiming at the problem of unstable training of the DDPG algorithm,a rule-guided DDPG algorithm was proposed,and a dual-value network was added in the training process.The experimental results show that the proposed method can achieve the scheduling goal better,and improve the stability and performance of the model on the original basis.
作者
黄尽云
罗倩
成梁成
HUANG Jinyun;LUO Qian;CHENG Liangcheng(School of Information and Communication Engineering,Beijing Information Science&Technology University,Beijing 100101,China;MOE Key Laboratory for Optoelectronic Measurement Technology and Instrument,Beijing Information Science&Technology University,Beijing 100101,China;China Electric Power Research Institute,Beijing 100192,China)
出处
《北京信息科技大学学报（自然科学版）》
2022年第2期56-61,共6页
Journal of Beijing Information Science and Technology University
关键词
深度确定性策略梯度
规则引导函数
电网调度
多目标
双价值网络
新能源消纳
deep deterministic policy gradient(DDPG)
rule-guiding function
grid dispatch
multi-objective
double-value network
new energy consumption
