摘要
从两个方面讨论具有最小二乘谱约束的对称斜哈密尔顿矩阵的逼近问题:(Ⅰ)研究使AX-XA的Frobenius范数最小的n阶实对称斜哈密尔顿矩阵A的集合L,其中X,A分别是特征向量和特征值矩阵,(Ⅱ)求A∈L使得‖C-A‖=(?)‖C-A‖,这里‖·‖是Frobenius范数.给出了L的元素的一般表达式和A的显示表达式,分析了该最佳逼近矩阵A的扰动理论,并给出了数值实验.
A nearness matrix problem is considered with two constraints--least square spectra constraint, symmetric and skew-Hamiltonian structure. It discusses two problems: (Ⅰ) the set L of symmetric and skew-Hamiltonian real n × n matrices A to minimize the Frobenius norm of AX - XA, where X, A are eigenvector and eigenvalue matrices, respectively, and (Ⅱ) find A∈L such that ‖C - A‖=min_ A∈L‖C - A‖, where ‖·‖ is the Frobenius norm. A general form of elements in L is given and an explicit expression of the minimizer A is derived. Perturbation theory of the nearest matrix is analyzed. A numerical example is reported.
出处
《数学物理学报(A辑)》
CSCD
北大核心
2013年第1期37-45,共9页
Acta Mathematica Scientia
基金
北京市自然科学基金(1122015)
北京市属高等学校人才强教深化计划项目(PHR201006116)资助
关键词
最佳逼近
最小二乘问题
扰动理论
Best approximation
Least squares problem
Perturbation theory.