用遗传算法求解有理函数模型  

Solving Rational Function Model with Genetic Algorithm

作　　者：祝汶琪[1] 焦伟利[1] 

机构地区：[1]中国科学院对地观测与数字地球科学中心

出　　处：《科学技术与工程》2008年第13期3530-3535,共6页Science Technology and Engineering

基　　金：国家863项目(2006AA122118)资助

摘　　要：传统的有理函数模型(RFM,Rational Function Model)参数的求解方法通常采用最小二乘法,在模型求解过程中,由于法方程病态,使求解结果很不稳定。针对这一问题,提出用具有搜索优秀结果能力的遗传算法求解有理函数模型的思想。实验结果表明,遗传算法可以有效地解决RFM求解过程中遇到的病态问题,并可求得所需的有理函数模型的精确解。The traditional method of solving the parameters of rational function model is least squares technique. In the solving process the ill-conditioned normal equation makes the results unstable. To solve the problem a new idea of solving rational function model has been put forward by using genetic algorithm which has the ability of searching excellent results in this paper. The experimental results indicate that genetic algorithm can solve ill-conditioned problem effectively and get the exact solution of rational function model.

关 键 词：有理函数模型 遗传算法 最小二乘法 病态方程 

分 类 号：TP391.75[自动化与计算机技术—计算机应用技术]