基于组合遗传神经网络的磨损趋势预测

Prediction of Wear Trend Using Combined BP Neural Network Based on Genetic Algorithm

针对神经网络易陷入局部极值的问题,用遗传算法来改进神经网络时间序列预测模型,并设计了一个组合网络,考虑了单一遗传BP神经网络预测的误差,以误差样本训练了一个校正的遗传BP神经网络,并将该组合网络时间序列预测模型应用于柴油机系统磨损趋势的预测,取得了较好的预测效果。

For the problem that BP neural network easily gets into the local extremism, the optimization of the structure and parameter of BP neural network using genetic algorithm was presented, and the combined network was developed of constructing auxiliary network to take the deviations of a single network into consideration. This combined network was used to forecast the wear trend of diesel system. The results are very satisfactory.