【作者】 张军； 王秋萍； 马生昀； 吕雄； 姚贵平；

【Author】 Jun Zhang;Qiuping Wang;Sheng-yun Ma;Xiong Lv;Guiping Yao;School of Science, Inner Mongolia Agricultural University;School of Science, Xi’anUniversity of Technology;

【机构】 内蒙古农业大学理学院； 西安理工大学理学院；

【摘要】 根据函数变换理论和灰色系统建模理论,本文引入了含有三个参数的函数变换,建立了以平均相对误差最小为目标的三参数函数变换GM(1,1)模型,同时运用遗传算法计算变换函数中的参数,实际例子的计算结果表明本文提出的函数变换满足光滑比缩小、级比偏差缩小、还原误差不扩大的条件,经过三参数函数变换改进的灰色预测模型的拟合效果和预测精度均有很大的提高。更多还原

【Abstract】 According to the function transformation theory and grey system modeling theory, function transformation with three parameters was proposed in this paper. Three parameters function transformation GM(1,1) model based on minimizing average relative error was build up, with genetic algorithm calculating the parameters of transformation functions. The calculation results of actual example show that function transformation proposed in this paper met smooth ratio lessening, stepwise ratio deviation lessening, error of inverse transformation without expanding. And the fitting and forecasting precision of three parameter function transformation GM(1,1) model are greatly improved in this paper.更多还原