摘要
鉴于航材备件在飞机维修保障中的重要地位,利用神经网络能以任意精度接近非线性函数的特点,采用神经网络的有监督式学习,并将预测误差作为反馈来调整航材备件需求率预测网络中的权值分布。其预测网络中只取了3个输入,增加输入个数还可使神经网络的学习效果更佳,从而使预测精度更高。通过对实际数据的预测结果表明,该算法具有较好的实用性。
Aiming at the importance of aviation spare parts in plane maintenance and safeguard, utilize characteristics that neural network can approach nonlinear function with arbitrary precision, adopted neural network's learning method with supervision, and regarded the prediction error as feedback to adjust the weighted values in aviation material spare requirement rate prediction network in order to achieve the objective of study. It has three importations in the prediction network, if increase the number of importation the study effect of neural network and the estimate precision will be better. The estimate result of a practicality suggests that the algorithm has good practicability.
出处
《兵工自动化》
2009年第1期54-55,64,共3页
Ordnance Industry Automation
关键词
备件需求
神经网络
BP算法
预测分析
Spares requirement
Neural network
BP algorithm
Estimate analysis