基于AGA—RBF神经网络的短期负荷预测  

Short-Term Load Forecasting Based on AGA-RBF Neural Network

作　　者：黄永高[1] 卢毅[1] 

机构地区：[1]东南大学电气工程学院,江苏南京210096

出　　处：《江苏电机工程》2009年第6期68-70,共3页Jiangsu Electrical Engineering

摘　　要：针对RBF神经网络的不足以及传统遗传算法的特点，采用浮点数编码的自适应遗传算法（AGA）作为RBF神经网络的学习算法．来确定RBF神经网络的隐含层的中心参数和宽度参数，形成AGA—RBF网络来进行负荷预测，并通过实例验证，该方法与RBF神经网络相比，能有效地提高预测精度和改善网络性能。In view of the insufficiency of the RBF neural network and the characteristic of the tradition genetic algorithm, the adaptive genetic algorithm encoded by floating-point numbers is used as a learning algorithm for RBF neural network to determine the central parameters and the width parameters of the hidden layers. The AGA-RBF neural network is used for short term load forecasting. The results of a practical example demonstrate that the AGA-RBF neural network can effectively improve the forecast accuracy and improve the network performance compared with the RBF neural network.

关 键 词：短期电力负荷预测 RBF神经网络 自适应遗传算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]