基于粗糙集遗传神经网络短期负荷预测方法_徐剑.pdfVIP

36 6 Vol.36 No.6 2008 3 16 RELAY Mar.16, 2008 Short-term load forecasting method of genetic algorithm neural network based on rough set XU Jian, LU Jun, SUN Yi, TANG Liang-rui (College of Electric and Electronic Engineering, North China Electric Power University, Beijing 102206, China) Abstract: With advantages of rough set, genetic algorithm and neural network, the paper proposes a new short-term load forecasting methodgenetic algorithm neural network based on rough set. Due to a mass of influencing short-term load forecasting factors, by using attribute algorithm of rough set theory is introduced to choose neural network’s input parameters. Parameters with a high correlation are used for input, which reduce the work and calculation time. In order to solve the shortcoming in the BP algorithm, such as slowness in training speed and convergence to the local minimum, genetic algorithm with the ability of strong global search is integrated. Forecasting results of calculation examples show that the new load forecasting method improve the speed and the accuracy of prediction, which is feasible and effective in the short-term load forecasting. Key words: short-term load forecasting; rough set; neural network; genetic algorithm TM715 A 1003-4897(2008)06-0037-05 BP