Short-term power load prediction method based on IGWO-Attention-GRU

The invention relates to the field of short-term power load prediction. According to the IGWO-Attention-GRU-based short-term power load prediction method, aiming at the defects of GWO, population initialization is completed by adopting an elite reverse learning strategy, and a population position is updated through a random differential variation mode. According to the method, an improved grey wolf optimization algorithm (IGWO) is used for optimizing Attention-GRU, and the short-term power load is predicted. According to the method, the hyperparameters of the Attention-GRU are searched by adopting IGWO, and the hyperparameters are preferentially obtained by a computer and do not need to be manually adjusted, so that the phenomenon of error increase caused by improper manual setting of the hyperparameters is avoided, and the precision of short-term power load prediction is effectively improved. 本发明涉及短期电力负荷预测领域。基于IGWO-Attention-GRU的短期电力负荷预测方法,针对GWO存在的不足，采用精英反向学习的策略完成种群初始化，并通过随机差分变异的方式更新种群位置。使用改进的灰狼优化算法（Improved