人工神经元网络负荷静态特性辨识的并行算法

THE PARALLEL ALGORITHM OF THE STATIC CHARACTERISTIC IDENTIFICATION OF ELECTRIC LOADS VIA ARTIFICATION NEURAL NETWORKS

基于人工神经元网络理论，对概率学习算法（ＰｒｏｂａｂｉｌｉｓｔｉｃＬｅａｒｎｉｎｇ）和误差反向传播算法（简称ＢＰ算法）的电力系统负荷静态特性辨识的并行算法进行了研究，并对这两种算法的计算复杂性和通信复杂性以及并行效率进行了分析。结果表明，这两种算法都有较好的并行性能。

In this paper ,the parallel algorithm of the static characteristic identification of electric loads based on Artificial Neural Networks(ANN)is developed by applying Probabilistic Learning and Error Back Propagation ( BP Algorithm). The computation complexity,communication complexity and the parallel effi-ciency of these two methods are given. The results show that both of these two parallel algorithms have good performance.