基于神经网络灰色Verhulst算法的装备研制费用预测模型

为了提高装备寿命周期费用LCC(life cycle cost)中研制费用的预测精度,对于呈S型的研制费用累加序列,利用灰色Verhulst模型对其预测.构造基于时间序列的人工神经网络输入输出模式对,利用BP神经网络对灰色Verhulst预测结果进行训练.仿真实例表明,该组合算法预测结果比单纯使用灰色Verhulst模型所得到的结果总体误差要小,将神经网络引入到灰色Verhulst模型中能较好地提高预测精度.

Forecasting model for development cost of equipment based on artificial neural network and grey Verhulst algorithm

In o rder to enhance the forecast precisio n of development cost in equipment life cycle co st ( LCC) , g rey Verhulst model w as applied to for ecast ing dev elo pment cost accumulat ion sequence w ith the shape of alphabet S. Couples of input and output data of art ificial neural netw ork based on t ime sequence w er e formed, and BP neural netw ork utilized to t rain the g rey forecast result . The simulat ion example show s that the hybrid algor ithm has few er erro rs than grey Verhulst mo del. The int roductio n of BP netw ork into g rey Verhul st model advances the forecast precision.