Scholarlink

Power load forecasting plays an important role in power system planning and operation, and is of key significance to ensure the stability and economy of power supply. This paper aims to improve the traditional power load forecasting methods, proposes an improved Ant-Lion Optimization (ALO) algorithm, and uses it to optimize the Long Short-Term Memory network (LSTM) model. First, the existing data set is cleaned. Secondly, the correlation of various factors in the data set is analyzed, and the data with high correlation with power load is retained to improve the accuracy of the model and reduce the running time. Then, the improved ALO algorithm is combined with the LSTM model to construct the improved ALO-LSTM model. Finally, LSTM and other models are used as comparative experiments to verify the feasibility of this model. Compared with ARIMA, LSTM, GA-LSTM, ALO-LSTM and models with attention mechanism, MAE decreased by 198.86, 16.44, 23.27, 41.05 and 114.52, respectively. RMSE decreased by 250.6, 17.96, 26.33, 45.15 and 146.08, respectively. The experimental results show that the improved ALO-LSTM model proposed in this study has potential application prospects in the field of power load forecasting.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Research on Improved Ant-Lion Optimization Long Short-Term Memory Network Power Load Forecasting Model

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jianjun Xu,
  • Yuanbo Shi,
  • Yueyang Huang

摘要

Power load forecasting plays an important role in power system planning and operation, and is of key significance to ensure the stability and economy of power supply. This paper aims to improve the traditional power load forecasting methods, proposes an improved Ant-Lion Optimization (ALO) algorithm, and uses it to optimize the Long Short-Term Memory network (LSTM) model. First, the existing data set is cleaned. Secondly, the correlation of various factors in the data set is analyzed, and the data with high correlation with power load is retained to improve the accuracy of the model and reduce the running time. Then, the improved ALO algorithm is combined with the LSTM model to construct the improved ALO-LSTM model. Finally, LSTM and other models are used as comparative experiments to verify the feasibility of this model. Compared with ARIMA, LSTM, GA-LSTM, ALO-LSTM and models with attention mechanism, MAE decreased by 198.86, 16.44, 23.27, 41.05 and 114.52, respectively. RMSE decreased by 250.6, 17.96, 26.33, 45.15 and 146.08, respectively. The experimental results show that the improved ALO-LSTM model proposed in this study has potential application prospects in the field of power load forecasting.