基于离散指数函数优化GM(1,1)模型的重新优化

Re-optimizing an Optimal GM(1,1) Model Based on Discrete Function with Exponential Law

分析了已存两种基于离散指数函数优化的GM(1,1)模型存在的问题,结合两者的优化思想对已存的背景值构造形式添加动态修正项,建立了适合于近似指数序列建模的重新优化模型,实例应用结果显示重新优化的模型大幅度提高了已存优化模型的精度.

The defects of two existing GM（1,1） models based on the discrete function with exponential law are analyzed. Dynamic correction terms are added to the existing construction of background value based on the optimization ideas of the two models. So a re-optimized model which is propitious to approximate non-ho- mogenous exponential sequence modeling is built. A case study shows that the re-optimized GM（1,1） model can improve the forecasting precision obviously.