基于小波变换和GM-ARMA的导弹备件消耗预测

针对导弹备件消耗呈现"小样本、非平稳"的特点, 为了克服传统预测方法依靠大样本数据进行建模的不足,提出了把基于小波变换和改进GM-ARMA的组合预测方法应用于导弹备件消耗预测的构想.在利用小波分解和其他模型建立组合模型的过程中,提出了先对小波基方程和分解层数2个特征进行参数化,再定量地对所有子模型的特征参数进行统一、综合的评估,以达到建立最佳组合模型的目的;然后对具有平稳特性的高频信息用阻尼最小二乘法优化的ARMA(Autoregressive and Moving Average)模型进行预测,对反映整体趋势体现非平稳的低频信息用背景值优化和数据变换技术改进的GM(1,1)模型进行预测.实例结果表明所提出的组合预测方法大大降低了预测误差,说明了该方法的有效性、可行性和实用性.

Consumption forecasting of missile spare parts based on wavelet transform and revised GM-ARMA model

Aiming at the characteristic of missile spare parts consumption presenting "small sample and non-stationary", and to overcome the deficiency that traditional forecasting method builds models with abundant date sample,the combined forecast method based on wavelet transform and revised GM-ARMA was proposed to carry through consumption forecasting of missile spare parts. In the process of building a hybrid model by combining wavelet decomposition with other forecasting model,first the new parameters of wavelet functions and decomposition levels were introduced, and all the parameters as a whole for the purpose of building an optimal hybrid model were quantitatively estimated. Then the high frequency signals were forecasted with ARMA model optimized by damping least-squares method, and the low frequency was forecasted with improved GM(1,1) model based on background value optimization and transformation. Experimental results show that the combined forecasting method significantly reduces forecast errors and is valid, feasible and useful in terms of forecasting missile spare parts consumption.