摘要
利用"最小二乘法"确定GOM(1,1)模型白化权函数的时间响应函数中的常数c,摈弃了传统GOM(1,1)模型将x(1)(1)作为初始值的欠科学的做法,构建了时间响应函数的优化模型。不管GOM(1,1)模型的发展系数a及控制系数b的估计值使用何种方法求出,通过本文两个准则确定待定系数c值从而得到时间响应函数的优化模型均能提高模型的模拟预测精度,并用实例验证了新方法的有效性与实用性。
Instead of usingx(1) ( 1 ) as the initial value for traditional GOM ( 1, 1 ) model, in this pa- per we utilize the method of "the least square estimate" to find the constant number c in the time re- sponse sequence of whiterization equation of GOM ( 1,1 ) , and create the optimum time response se- quence for GOM ( 1,1 ). No matter what method can be used to obtain development coefficient a and control coefficient b of GOM ( 1,1 ) , we can always get the optimum time response sequence for GOM ( 1,1 ) by using the two criteria in this paper to find the constant number c. and proved its reasonabili- ty and feasibility with two example.
出处
《贵州师范大学学报(自然科学版)》
CAS
2013年第6期66-69,共4页
Journal of Guizhou Normal University:Natural Sciences
基金
西华师范大学青年资助专项项目(130018)
作者简介
万琴(1983-),女,四川省绵阳市人,硕士、讲师.研究方向:灰色系统理论及其应用.wanqin1014@126.com