前馈神经网络的模糊 PID 算法及其在电力系统负荷预测中的应用

The Fuzzy PID Algorithm for Feedforward Neural Network and its Application in Electric Power System Load Forecasting

提出了多层前馈神经网络的模糊ＰＩＤ学习算法（ＦＰＢＰ）。这种算法是把多层前馈神经网络的学习过程当作一个动态控制系统来处理，确定出动态控制系统达到稳态时的ＰＩＤ控制器参数，然后再基于模糊控制的思想，对确定出的ＰＩＤ控制器参数进行模糊调整。文中给出了这种算法在电力系统负荷预测中的实际应用，并与标准ＢＰ算法作了比较。结果表明，该算法提高了网络的学习速度和预测的精度。

The Fuzzy PID Back Propagation (FPBP)algorithm for multilayer feedforward neural network is proposed in this paper.The algorithm deals the learning process of multilayer feedforward neural network as a dynamic control system and introducesthe PID controller into the dynamic control system.Then adjusts the PID controller parameters by the concept of fuzzy control.An example of the algorithm's application in electric power system load forecasting is shown and compared with standard BP algorithm.The result indicats that the algorithm increases the networklearning speed and the accuracy.