基于GM-RBF神经网络的导弹武器系统使用可用度评估方法 研究

目的研究导弹武器系统使用可用度评估问题，方法通过基于故障数据的使用可用度评估，提出一种基于灰色模型GM（1，1）的径向基函数（RBF—Radial Basis Function）神经网络组合模型。结果该模型克服了灰色理论的长时间序列预估误差大和神经网络的训练样本需求量大、输入变量选取困难等缺点。结论仿真结果表明，GM—RBF神经网络对导弹武器系统使用可用度评估具有评估误差小、精度高等优点。

Study on the Operational Availability Evaluation Method of Missile Weapon System Based on GM-RBF Neural Network

Objective To investigate the operational availability evaluation method of missile weapon system. Methods According to the operational availability evaluation problem of missile weapon system, a combined model was put forward by operational availability evaluation using fault data based on GM （1,1） and RBF neural network. Results The model overcame the disadvantages of big error and long time series prediction of GM （1,1） and great demand of training samples and difficulty of variables selection of neural network. Conclusion Experimental results reveal that the operational availability evaluation of missile weapon system by GM-RBF neural network owns the advantages of evaluation minor error and high precision.