Namespaces
Variants
Actions

Difference between revisions of "Cholesky factorization"

From Encyclopedia of Mathematics
Jump to: navigation, search
m (fix)
m (Automatically changed introduction)
 
(2 intermediate revisions by the same user not shown)
Line 1: Line 1:
A symmetric <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c1201601.png" /> matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c1201602.png" /> (cf. also [[Symmetric matrix|Symmetric matrix]]) is positive definite if the [[Quadratic form|quadratic form]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c1201603.png" /> is positive for all non-zero vectors <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c1201604.png" /> or, equivalently, if all the eigenvalues of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c1201605.png" /> are positive. Positive-definite matrices have many important properties, not least that they can be expressed in the form <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c1201606.png" /> for a non-singular matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c1201607.png" />. The Cholesky factorization is a particular form of this factorization in which <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c1201608.png" /> is upper triangular with positive diagonal elements, and it is usually written as <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c1201609.png" />. In the case of a scalar (<img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016010.png" />), Cholesky factorization corresponds to the fact that a positive number has a positive square root. The factorization is named after A.-L. Cholesky, a French military officer involved in geodesy. It is closely connected with the solution of least-squares problems (cf. also [[Least squares, method of|Least squares, method of]]), since the normal equations that characterize the least-squares solution have a symmetric positive-definite coefficient matrix.
+
<!--This article has been texified automatically. Since there was no Nroff source code for this article,
 +
the semi-automatic procedure described at https://encyclopediaofmath.org/wiki/User:Maximilian_Janisch/latexlist
 +
was used.
 +
If the TeX and formula formatting is correct and if all png images have been replaced by TeX code, please remove this message and the {{TEX|semi-auto}} category.
  
The Cholesky factorization can be computed by a form of [[Gaussian elimination]] that takes advantage of the symmetry and definiteness. Equating <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016011.png" /> elements in the equation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016012.png" /> gives
+
Out of 47 formulas, 46 were replaced by TEX code.-->
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016013.png" /></td> </tr></table>
+
{{TEX|semi-auto}}{{TEX|part}}
 +
A symmetric $( n \times n )$ matrix $A$ (cf. also [[Symmetric matrix|Symmetric matrix]]) is positive definite if the [[Quadratic form|quadratic form]] $x ^ { T } A x$ is positive for all non-zero vectors $x$ or, equivalently, if all the eigenvalues of $A$ are positive. Positive-definite matrices have many important properties, not least that they can be expressed in the form $A = X ^ { T } X$ for a non-singular matrix $X$. The Cholesky factorization is a particular form of this factorization in which $X$ is upper triangular with positive diagonal elements, and it is usually written as $A = R ^ { T } R$. In the case of a scalar ($n = 1$), Cholesky factorization corresponds to the fact that a positive number has a positive square root. The factorization is named after A.-L. Cholesky, a French military officer involved in geodesy. It is closely connected with the solution of least-squares problems (cf. also [[Least squares, method of|Least squares, method of]]), since the normal equations that characterize the least-squares solution have a symmetric positive-definite coefficient matrix.
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016014.png" /></td> </tr></table>
+
The Cholesky factorization can be computed by a form of [[Gaussian elimination]] that takes advantage of the symmetry and definiteness. Equating $( i , j )$ elements in the equation $A = R ^ { T } R$ gives
  
These equations can be solved to yield <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016015.png" /> a column at a time, according to the following algorithm:
+
\begin{equation*} j = i :\, a _ { i i } = \sum _ { k = 1 } ^ { i } r _ { k i } ^ { 2 }, \end{equation*}
  
<table border="0" cellpadding="0" cellspacing="0" style="background-color:black;"> <tr><td> <table border="0" cellspacing="1" cellpadding="4" style="background-color:black;"> <tbody> <tr> <td colname="1" style="background-color:white;" colspan="1">FOR <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016016.png" /></td> </tr> <tr> <td colname="1" style="background-color:white;" colspan="1">FOR <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016017.png" /></td> </tr> <tr> <td colname="1" style="background-color:white;" colspan="1">  <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016018.png" /></td> </tr> <tr> <td colname="1" style="background-color:white;" colspan="1">END</td> </tr> <tr> <td colname="1" style="background-color:white;" colspan="1"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016019.png" /></td> </tr> <tr> <td colname="1" style="background-color:white;" colspan="1">END</td> </tr> </tbody> </table>
+
\begin{equation*} j &gt; i : a _ { ij } = \sum _ { k = 1 } ^ { i } r _ { k i } r _ { k j }. \end{equation*}
 +
 
 +
These equations can be solved to yield $R$ a column at a time, according to the following algorithm:
 +
 
 +
<table border="0" cellpadding="0" cellspacing="0" style="background-color:black;"> <tr><td> <table border="0" cellpadding="4" cellspacing="1" style="background-color:black;"> <tr> <td colname="1" colspan="1" style="background-color:white;">FOR $j = 1 : n$</td> </tr> <tr> <td colname="1" colspan="1" style="background-color:white;">FOR $i = 1 : j - 1$</td> </tr> <tr> <td colname="1" colspan="1" style="background-color:white;">  <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016018.png"/></td> </tr> <tr> <td colname="1" colspan="1" style="background-color:white;">END</td> </tr> <tr> <td colname="1" colspan="1" style="background-color:white;">$r _ { j j } = \left( a _ { j j } - \sum _ { k = 1 } ^ { j - 1 } r _ { k j } ^ { 2 } \right) ^ { 1 / 2 }$</td> </tr> <tr> <td colname="1" colspan="1" style="background-color:white;">END</td> </tr> </table>
  
 
</td></tr> </table>
 
</td></tr> </table>
  
It is the positive definiteness of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016020.png" /> that guarantees that the argument of the square root in this algorithm is always positive.
+
It is the positive definiteness of $A$ that guarantees that the argument of the square root in this algorithm is always positive.
  
It can be shown [[#References|[a1]]] that in floating-point arithmetic the computed solution <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016021.png" /> satisfies <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016022.png" />, where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016023.png" />, with <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016024.png" /> a constant of order <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016025.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016026.png" /> the unit round-off (or machine precision). Here, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016027.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016028.png" />. This excellent numerical stability is essentially due to the equality <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016029.png" />, which guarantees that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016030.png" /> is of bounded norm relative to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016031.png" />.
+
It can be shown [[#References|[a1]]] that in floating-point arithmetic the computed solution $\hat{x}$ satisfies $( A + \Delta A ) \hat{x} = b$, where $\| \Delta A \| _ { 2 } \leq c n ^ { 2 } u \| A \| _ { 2 }$, with $c$ a constant of order $1$ and $u$ the unit round-off (or machine precision). Here, $\|A \| _ { 2 } = \operatorname { max } _ { x \neq 0} \|Ax\|_2 / \| x \|_2$ and $\| x \| _ { 2 } = ( x ^ { T } x ) ^ { 1 / 2 }$. This excellent numerical stability is essentially due to the equality $\| A \| _ { 2 } = \| R ^ { T } R \| _ { 2 } = \| R \| _ { 2 } ^ { 2 }$, which guarantees that $R$ is of bounded norm relative to $A$.
  
A variant of Cholesky factorization is the factorization <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016032.png" /> where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016033.png" /> is unit lower triangular and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016034.png" /> is diagonal. This factorization exists for definite matrices and some (but not all) indefinite ones. When <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016035.png" /> is positive definite, the Cholesky factor is given by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016036.png" />.
+
A variant of Cholesky factorization is the factorization $A = L D L ^ { T }$ where $L$ is unit lower triangular and $D$ is diagonal. This factorization exists for definite matrices and some (but not all) indefinite ones. When $A$ is positive definite, the Cholesky factor is given by $R = D ^ { 1 / 2 } L ^ { T }$.
  
A symmetric matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016037.png" /> is positive semi-definite if the quadratic form <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016038.png" /> is non-negative for all <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016039.png" />; thus, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016040.png" /> may be singular (cf. also [[Matrix|Matrix]]). For such matrices a Cholesky factorization exists, but <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016041.png" /> may not display the rank of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016042.png" />. However, by introducing row and column permutations it is always possible to obtain a factorization
+
A symmetric matrix $A$ is positive semi-definite if the quadratic form $x ^ { T } A x$ is non-negative for all $x$; thus, $A$ may be singular (cf. also [[Matrix|Matrix]]). For such matrices a Cholesky factorization exists, but $R$ may not display the rank of $A$. However, by introducing row and column permutations it is always possible to obtain a factorization
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016043.png" /></td> </tr></table>
+
\begin{equation*} \Pi ^ { T } A \Pi = R ^ { T } R , \quad R = \left( \begin{array} { c c } { R _ { 11 } } &amp; { R _ { 12 } } \\ { 0 } &amp; { 0 } \end{array} \right), \end{equation*}
  
where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016044.png" /> is a permutation matrix, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016045.png" /> is <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016046.png" /> upper triangular with positive diagonal elements, and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c120/c120160/c12016047.png" />.
+
where $\Pi$ is a permutation matrix, $R _ { 11 }$ is $( r \times r )$ upper triangular with positive diagonal elements, and $\operatorname {rank} ( A ) = r$.
  
 
Cf. [[Matrix factorization|Matrix factorization]] for more details and references.
 
Cf. [[Matrix factorization|Matrix factorization]] for more details and references.
  
 
====References====
 
====References====
<table><TR><TD valign="top">[a1]</TD> <TD valign="top">  N.J. Higham,  "Accuracy and stability of numerical algorithms" , SIAM (Soc. Industrial Applied Math.)  (1996)</TD></TR></table>
+
<table><tr><td valign="top">[a1]</td> <td valign="top">  N.J. Higham,  "Accuracy and stability of numerical algorithms" , SIAM (Soc. Industrial Applied Math.)  (1996)</td></tr></table>

Latest revision as of 17:44, 1 July 2020

A symmetric $( n \times n )$ matrix $A$ (cf. also Symmetric matrix) is positive definite if the quadratic form $x ^ { T } A x$ is positive for all non-zero vectors $x$ or, equivalently, if all the eigenvalues of $A$ are positive. Positive-definite matrices have many important properties, not least that they can be expressed in the form $A = X ^ { T } X$ for a non-singular matrix $X$. The Cholesky factorization is a particular form of this factorization in which $X$ is upper triangular with positive diagonal elements, and it is usually written as $A = R ^ { T } R$. In the case of a scalar ($n = 1$), Cholesky factorization corresponds to the fact that a positive number has a positive square root. The factorization is named after A.-L. Cholesky, a French military officer involved in geodesy. It is closely connected with the solution of least-squares problems (cf. also Least squares, method of), since the normal equations that characterize the least-squares solution have a symmetric positive-definite coefficient matrix.

The Cholesky factorization can be computed by a form of Gaussian elimination that takes advantage of the symmetry and definiteness. Equating $( i , j )$ elements in the equation $A = R ^ { T } R$ gives

\begin{equation*} j = i :\, a _ { i i } = \sum _ { k = 1 } ^ { i } r _ { k i } ^ { 2 }, \end{equation*}

\begin{equation*} j > i : a _ { ij } = \sum _ { k = 1 } ^ { i } r _ { k i } r _ { k j }. \end{equation*}

These equations can be solved to yield $R$ a column at a time, according to the following algorithm:

FOR $j = 1 : n$
FOR $i = 1 : j - 1$
END
$r _ { j j } = \left( a _ { j j } - \sum _ { k = 1 } ^ { j - 1 } r _ { k j } ^ { 2 } \right) ^ { 1 / 2 }$
END

It is the positive definiteness of $A$ that guarantees that the argument of the square root in this algorithm is always positive.

It can be shown [a1] that in floating-point arithmetic the computed solution $\hat{x}$ satisfies $( A + \Delta A ) \hat{x} = b$, where $\| \Delta A \| _ { 2 } \leq c n ^ { 2 } u \| A \| _ { 2 }$, with $c$ a constant of order $1$ and $u$ the unit round-off (or machine precision). Here, $\|A \| _ { 2 } = \operatorname { max } _ { x \neq 0} \|Ax\|_2 / \| x \|_2$ and $\| x \| _ { 2 } = ( x ^ { T } x ) ^ { 1 / 2 }$. This excellent numerical stability is essentially due to the equality $\| A \| _ { 2 } = \| R ^ { T } R \| _ { 2 } = \| R \| _ { 2 } ^ { 2 }$, which guarantees that $R$ is of bounded norm relative to $A$.

A variant of Cholesky factorization is the factorization $A = L D L ^ { T }$ where $L$ is unit lower triangular and $D$ is diagonal. This factorization exists for definite matrices and some (but not all) indefinite ones. When $A$ is positive definite, the Cholesky factor is given by $R = D ^ { 1 / 2 } L ^ { T }$.

A symmetric matrix $A$ is positive semi-definite if the quadratic form $x ^ { T } A x$ is non-negative for all $x$; thus, $A$ may be singular (cf. also Matrix). For such matrices a Cholesky factorization exists, but $R$ may not display the rank of $A$. However, by introducing row and column permutations it is always possible to obtain a factorization

\begin{equation*} \Pi ^ { T } A \Pi = R ^ { T } R , \quad R = \left( \begin{array} { c c } { R _ { 11 } } & { R _ { 12 } } \\ { 0 } & { 0 } \end{array} \right), \end{equation*}

where $\Pi$ is a permutation matrix, $R _ { 11 }$ is $( r \times r )$ upper triangular with positive diagonal elements, and $\operatorname {rank} ( A ) = r$.

Cf. Matrix factorization for more details and references.

References

[a1] N.J. Higham, "Accuracy and stability of numerical algorithms" , SIAM (Soc. Industrial Applied Math.) (1996)
How to Cite This Entry:
Cholesky factorization. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Cholesky_factorization&oldid=37467
This article was adapted from an original article by N.J. Higham (originator), which appeared in Encyclopedia of Mathematics - ISBN 1402006098. See original article