基于DE-GWO-SVR的中长期电力需求预测

电力需求预测是电力系统科学规划与运行的重要前提。根据相关性分析，从经济发展水平、城镇化水平、工业化水平、人口数量、产业结构、居民消费水平、电价和用电基数8个方面筛选出电力需求关键影响因素。利用差分进化（differential evolution，DE）和灰狼优化（grey wolf optimization，GWO）算法对支持向量回归模型（support vector regression，SVR）的参数进行优化，建立差分进化-灰狼优化-支持向量回归电力需求预测模型。选取北京市电力需求历史数据进行实证分析，比较不同模型的预测结果，验证组合优化模型的有效性及其预测的准确率，并对北京市2021—2025年电力需求进行预测。

Medium and Long-Term Power Demand Forecasting Based on DE-GWO-SVR

Power demand forecasting is an important prerequisite for the scientific planning and operation of power systems. According to the results of the correlation analysis, the key influencing factors of power demand are selected from eight aspects: economic development level, urbanization level, industrialization level, population size, industrial structure, household consumption level, electricity price and electricity base. Using differential evolution (DE) and grey wolf optimization (GWO) algorithms to optimize the parameters of the support vector regression (SVR), the power demand forecasting model is established. Based on the historical data of power demand in Beijing, this paper makes an empirical analysis, compares the prediction results of different models, verifies the effectiveness of the combined optimization model and the accuracy of prediction, forecasts the power demand of Beijing from 2021 to 2025.