基于混沌学习算法的神经网络短期负荷预测

基于混沌理论对电力负荷的复杂时间序列进行分析，得出该时间序列属于混沌序列的结论，就此提出了一种新的神经网络（NN）学习算法——混沌学习算法。该算法中的混沌轨道的游动性有利于系统跳出局域极值的束缚而寻求全局最优，这样克服了前馈NN 的BP学习算法所存在的本质问题，使NN训练的收敛性好、速度快、误差小。文中通过对实际系统负荷预测结果，与BP算法预测结果比较，证明了混沌学习算法的电力负荷短期预测具有明显好的效果。

SHORT-TERM LOAD FORECASTING USING CHAOTIC LEARNING ALGORITHM FOR NEURAL NETWORK

This paper analyzes the time series of power load using chaotic theory. then concludes that power load time seriesbelongs to chaotic series.' A new neural network (NN) learning algorithm, chaotic learning algorithm, is proposed in thispaper. This algorithm has similar ability with random search approach so that the walking of chaotic track can escape fromthe local extremum and reach the global optimum. The actual forecasting results and comparison with the BP algorithm provethat chaotic learning algorithm has better forecasting accuracy and convergence.