基于粒子群优化算法的LightGBM 超短期负荷预测研究
周彬彬¹，蒋燕¹，赵珍玉¹，段睿钦¹，刘力铭²
（1. 云南电力调度控制中心，云南 昆明 650011；2. 北京清软创新科技股份有限公司，北京 100080）
摘要： 针对当前超短期负荷预测模型的不足，提出了一种基于粒子群优化算法的LightGBM 超短期负荷预测模型，实
现了LightGBM 模型参数的自适应调整，可针对不同时空下负荷的规律特点调整模型参数，提高了模型的可推广性。仿
真结果表明，提出的模型能有效提高负荷预测的准确度。
关键词： 粒子群优化算法；LightGBM；K-fold 交叉验证；超短期负荷预测
中图分类号： TM714     文献标识码： A     文章编号： 2095-0802-(2021)02-0002-05
Ultra-short Term Load Forecasting Research Based on Particle Swarm Optimization
Enhanced LightGBM
ZHOU Binbin¹, JIANG Yan¹, ZHAO Zhenyu¹, DUAN Ruiqin¹, LIU Liming²
(1. Yunnan Electric Power Dispatching Control Center, Kunming 650011, Yunnan, China; 2. Beijing Tsingsoft Technology Co.,
Ltd., Beijing 100080, China)
Abstract: Aiming at the deficiency of current ultra-short term load forecasting model, this paper proposed a utra-short term load
forecasting model based on particle swarm optimization(PSO) enhanced LightGBM. The model can realize the adaptive adjustment
of the parameters of LightGBM model. It can also adjust the model parameters according to the characteristics of the load in
different time and space, and improve the generalization of the model. The simulation results show that the proposed model can
improve the accuracy of load forecasting.

Key words: particle swarm optimization; LightGBM; K-fold cross-validation; ultra-short term load forecasting