Namespaces
Variants
Actions

Difference between revisions of "Oscillating matrix"

From Encyclopedia of Mathematics
Jump to: navigation, search
(Importing text file)
 
(TeX)
 
Line 1: Line 1:
 +
{{TEX|done}}
 
''oscillatory matrix''
 
''oscillatory matrix''
  
A totally non-negative [[Matrix|matrix]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o0704901.png" /> for which there exists a positive integer <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o0704902.png" /> such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o0704903.png" /> is a totally positive matrix; the matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o0704904.png" /> is called totally non-negative (totally positive) if all its minors, of whatever order, are non-negative (positive). The lowest exponent <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o0704905.png" /> is called the exponent of the oscillating matrix. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o0704906.png" /> is an oscillating matrix with exponent <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o0704907.png" />, then for any integer <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o0704908.png" /> the matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o0704909.png" /> is totally positive; an integer positive power of an oscillating matrix and the matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049010.png" /> are also oscillating matrices. In order that a totally non-negative matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049011.png" /> is an oscillating matrix, it is necessary and sufficient that: 1) <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049012.png" /> is a non-singular matrix; and 2) for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049013.png" />, the following are fulfilled: <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049014.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049015.png" />.
+
A totally non-negative [[Matrix|matrix]] $A$ for which there exists a positive integer $\chi$ such that $A^\chi$ is a totally positive matrix; the matrix $A$ is called totally non-negative (totally positive) if all its minors, of whatever order, are non-negative (positive). The lowest exponent $\chi$ is called the exponent of the oscillating matrix. If $A$ is an oscillating matrix with exponent $\chi$, then for any integer $k\geq\chi$ the matrix $A^k$ is totally positive; an integer positive power of an oscillating matrix and the matrix $(A^+)^{-1}$ are also oscillating matrices. In order that a totally non-negative matrix $A=\|a_{ik}\|_1^n$ is an oscillating matrix, it is necessary and sufficient that: 1) $A$ is a non-singular matrix; and 2) for $i=1,\dots,n$, the following are fulfilled: $a_{i,i+1}>0$, $a_{i+1,i}>0$.
  
The basic theorem on oscillating matrices is: An oscillating matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049016.png" /> always has <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049017.png" /> different positive eigen values; for the eigen vector <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049018.png" /> that corresponds to the largest eigen value <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049019.png" />, all coordinates differ from zero and are of the same sign; for an eigen vector <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049020.png" /> that corresponds to the <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049021.png" />-th eigen value <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049022.png" /> (arranged according to decreasing value) there are exactly <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049023.png" /> changes of sign; for any real numbers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049024.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049025.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049026.png" />, the number of changes of sign in the sequence of coordinates of the vector <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049027.png" /> is between <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049028.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/o/o070/o070490/o07049029.png" />.
+
The basic theorem on oscillating matrices is: An oscillating matrix $A=\|a_{ik}\|_1^n$ always has $n$ different positive eigen values; for the eigen vector $u^1$ that corresponds to the largest eigen value $\lambda_1$, all coordinates differ from zero and are of the same sign; for an eigen vector $u^s$ that corresponds to the $s$-th eigen value $\lambda_s$ (arranged according to decreasing value) there are exactly $s-1$ changes of sign; for any real numbers $c_g,\dots,c_h$, $1\leq g\leq h\leq n$, $\sum_{k=g}^hc_k^2>0$, the number of changes of sign in the sequence of coordinates of the vector $u=\sum_{k=g}^hc_ku^k$ is between $g-1$ and $h-1$.
  
 
====References====
 
====References====

Latest revision as of 16:31, 4 November 2014

oscillatory matrix

A totally non-negative matrix $A$ for which there exists a positive integer $\chi$ such that $A^\chi$ is a totally positive matrix; the matrix $A$ is called totally non-negative (totally positive) if all its minors, of whatever order, are non-negative (positive). The lowest exponent $\chi$ is called the exponent of the oscillating matrix. If $A$ is an oscillating matrix with exponent $\chi$, then for any integer $k\geq\chi$ the matrix $A^k$ is totally positive; an integer positive power of an oscillating matrix and the matrix $(A^+)^{-1}$ are also oscillating matrices. In order that a totally non-negative matrix $A=\|a_{ik}\|_1^n$ is an oscillating matrix, it is necessary and sufficient that: 1) $A$ is a non-singular matrix; and 2) for $i=1,\dots,n$, the following are fulfilled: $a_{i,i+1}>0$, $a_{i+1,i}>0$.

The basic theorem on oscillating matrices is: An oscillating matrix $A=\|a_{ik}\|_1^n$ always has $n$ different positive eigen values; for the eigen vector $u^1$ that corresponds to the largest eigen value $\lambda_1$, all coordinates differ from zero and are of the same sign; for an eigen vector $u^s$ that corresponds to the $s$-th eigen value $\lambda_s$ (arranged according to decreasing value) there are exactly $s-1$ changes of sign; for any real numbers $c_g,\dots,c_h$, $1\leq g\leq h\leq n$, $\sum_{k=g}^hc_k^2>0$, the number of changes of sign in the sequence of coordinates of the vector $u=\sum_{k=g}^hc_ku^k$ is between $g-1$ and $h-1$.

References

[1] F.R. Gantmakher, M.G. Krein, "Oscillation matrices and kernels and small vibrations of mechanical systems" , Dept. Commerce USA. Joint Publ. Service (1961) (Translated from Russian)


Comments

References

[a1] S. Karlin, "Total positivity" , Stanford Univ. Press (1960)
[a2] F.R. [F.R. Gantmakher] Gantmacher, "The theory of matrices" , 2 , Chelsea, reprint (1959) pp. Chapt. XIII, §9 (Translated from Russian)
How to Cite This Entry:
Oscillating matrix. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Oscillating_matrix&oldid=34294
This article was adapted from an original article by V.I. Lomonosov (originator), which appeared in Encyclopedia of Mathematics - ISBN 1402006098. See original article