摘要
对于非线性系统的直接加权优化辨识算法,通过在原线性仿射函数形式中增加若干项关于输入观测数据序列的线性项,来增强逼近非线性,减少逼近的时间.对于增加若干线性项后展开式中的未知权重值的选取,分别从理论和实用上推导出这些未知权重值的选取过程.理论上的推导分析,可明确增加的未知权重值在整个逼近非线性系统的目的中起着辅助作用;实用上的推导分析,将某些复杂的最优化问题经过整理变换成常见的最优化问题,从而可利用最为基础的优化方法来求解.
As for the direct weighted optimization identification algorithm in the nonlinear system,enhancing approximate nonlinear and reducing approximate time can be realized by increasing several items about the linear items of input observed sequence in the original linear affine-functions.As for the unknown weights selection in expansion equation after increasing several linear items,this paper deduced the selection process of these unknown weight value from the theory and practical angle respectively.Theoretical analysis can increase definitely auxiliary role of unknown weights in the purpose of approximate nonlinear system;On the practical analysis is showing how to transform the complicated optimization problems to the common optimization problems and finally the based optimization solution can be used to solve these problems.
出处
《华侨大学学报（自然科学版）》
CAS
北大核心
2013年第1期49-55,共7页
Journal of Huaqiao University(Natural Science)
基金
国家自然科学基金资助项目(61164014)
关键词
系统辨识
非线性系统
权重值
最优化问题
迭代
system identification
nonlinear system
weights value
optimization problems
iterative choose
