应用RBF网络的导弹研制费用预测  

Development Cost Predicting of RBF Neural Networks for Missile

作　　者：徐军凯[1] 李勇[1] 梁宇[1] 王俊涛[1] 

机构地区：[1]防空兵指挥学院研究生队,河南郑州450052

出　　处：《兵工自动化》2007年第10期55-56,59,共3页Ordnance Industry Automation

摘　　要：影响导弹武器系统研制费用的因素含最大有效射程、目标容量、导弹最大飞行速度、发射质量及研制周期。建立基于RBF神经网络的费用预测模型,其算法步骤包括:预测算法思想、归一化原始数据、样本集输入数据的选取。该模型不仅克服了因专家主观因素而造成的影响,且提高了算法的运行效率和模型的预测精度。The principal factors which affect the development cost of Missile weapon system contains the extreme validartillery range, target capacity, the highest airspeed, shoots quantity and development period. The model of predicting development cost is presented based on RBF neural networks, the program step of which contains predict the program thought, return on turning the original data, the selection in input data of the sample gathers. The model not only overcomes the negative effect of experts＇ subjectivity but also improves the process of the program and predicting precision.

关 键 词：导弹武器系统 研制费用 RBF神经网络 预测 

分 类 号：U661.3[交通运输工程—船舶及航道工程]