在具有实谱的矩阵大相空间上线性化Toda和SVD流
Linearizing Toda and SVD flows on large phase spaces of matrices with real spectrum
AI总结:
本文在具有实谱的矩阵大相空间上,仅利用线性代数基本事实构造图表,将Toda流和SVD流线性化为常系数线性微分系统或匀速直线运动,从而将渐近行为转化为局部问题。
摘要:
我们考虑Toda流和较不熟悉的SVD流的不同相空间。对于Toda流,我们处理具有实单特征值的对称和非对称矩阵,可能带有给定轮廓。例如,轮廓可以编码带状矩阵和Hessenberg矩阵。对于SVD流,我们假设奇异值具有简单性。在所有情况下,都构造了开覆盖,以及到欧几里得空间的相应图表。这些图表将流线性化,将其转换为具有常系数和对角矩阵的线性微分系统。一种变体构造将流转换为匀速直线运动。由于极限点属于相空间,渐近行为成为一个局部问题。这些构造仅依赖于线性代数的基本事实,不使用辛几何。
英文摘要:
We consider different phase spaces for the Toda flows and the less familiar SVD flows. For the Toda flow, we handle symmetric and non-symmetric matrices with real simple eigenvalues, possibly with a given profile. Profiles encode, for example, band matrices and Hessenberg matrices. For the SVD flow, we assume simplicity of the singular values. In all cases, an open cover is constructed, as are corresponding charts to Euclidean space. The charts linearize the flows, converting it into a linear differential system with constant coefficients and diagonal matrix. A variant construction transform the flows into uniform straight line motion. Since limit points belong to the phase space, asymptotic behavior becomes a local issue. The constructions rely only on basic facts of linear algebra, making no use of symplectic geometry.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。