基于遗传算法和径向基函数神经网络的短期边际电价预测  

Short-Term Marginal Price Forecasting Based on Genetic Algorithm and Radial Basis Function Neural Network

作　　者：顾庆雯[1] 陈刚[1] 朱蕾蕾[2] 吴迎霞[3] 

机构地区：[1]重庆大学电气工程学院,重庆市沙坪区400030 [2]浙江省湖州电力局,浙江省湖州市313000 [3]重庆电力调度通信中心,重庆市渝中区400014

出　　处：《电网技术》2006年第7期18-21,25,共5页Power System Technology

摘　　要：文章分析了影响电价的主要因素及电价的变化特点,讨论了电价预测模型中必需引入的影响电价的因素。在比较常用的几种电价预测方法的优缺点后,作者采用径向基函数神经网络(radial basis function neural networks,RBF)建立短期边际电价预测模型,用递阶遗传算法(HGA)同时训练RBF网络结构和参数。并以美国New England ISO公布的2002年历史电价数据进行训练和测试,与传统的BP网络预测模型相比较, 测试结果证明该模型的预测精确度是令人满意的。The main factors influencing electricity price and the variation features of electricity price are analyzed, the factors that must be led into electricity price forecasting model are researched. After comparing the advantages and defects of electricity price forecasting methods in common use, a short-term marginal price forecasting model is proposed by use of radial basis function （RBF） neural network, and the structure and parameters of RBF neural network are simultaneously trained by hierarchical genetic algorithm （HGA）. The proposed RBF neural network is trained and tested by historical data in the year of 2002 published by ISO of New England power grid. Test results show that the forecasted marginal electricity prices by the proposed model are satisfied and more accurate than those by traditional BP network.

关 键 词：短期边际电价 RBF网络 递阶遗传算法 电力市场 

分 类 号：TM715[电气工程—电力系统及自动化] F123.9[经济管理—世界经济]