灰色GM(1,1)模型的改进

Improvement of grey model GM(1,1)

在分析灰色GM(1,1)模型缺陷的基础上,导出了背景值构造法中权值p的精确表达式.由于GM(1,1)模型拟合式默认经过了初始点,与实际情况并不一定相符,因此在初始值中加入了扰动因子m,并以遗传算法求取了m值的大小.实例证明了模型的改进可以提高预测精度.

On the basis of analyzing the defect of grey model GM（1,1）,the accurate expression of weight p in the structure method of the background value is put forward.The simulated expression of GM（1,1） model is pretermitted to pass the initial point,which does not always happen in fact,so a disturbance factor m is introduced to the initial value and worked out by using genetic algorithm.Finally based on an example,it is proved that the improved model can increase the forecasting precision.