Difference between revisions of "Manuals/calci/CHOLESKY"

From ZCubes Wiki
Jump to navigation Jump to search
 
(29 intermediate revisions by 5 users not shown)
Line 1: Line 1:
−
<div style="font-size:30px">'''CHOLESKY(arr)'''</div><br/>
+
<div style="font-size:25px">'''CHOLESKY (Matrix) '''</div><br/>
−
*<math>arr</math> is the  array of numeric elements
+
*<math>Matrix</math> is the  array of numeric elements.
  
 
==Description==
 
==Description==
Line 7: Line 7:
 
*The Cholesky Factorization is only defined for symmetric or Hermitian positive definite matrices.
 
*The Cholesky Factorization is only defined for symmetric or Hermitian positive definite matrices.
 
*Every positive definite matrix A can be factored as  <math>A</math> = <math>LL^{T}</math>
 
*Every positive definite matrix A can be factored as  <math>A</math> = <math>LL^{T}</math>
−
where
 
 
  <math>L</math> is lower triangular with positive diagonal elements
 
  <math>L</math> is lower triangular with positive diagonal elements
 
  <math>L^{T}</math> is is the conjugate transpose value of <math>L</math>
 
  <math>L^{T}</math> is is the conjugate transpose value of <math>L</math>
 
*Every Hermitian positive-definite matrix  has a unique Cholesky decomposition.
 
*Every Hermitian positive-definite matrix  has a unique Cholesky decomposition.
−
*Here <math>CHOLESKY(arr)</math>, <math>arr</math> is set of values to find the factorization value.
+
*Here <math>CHOLESKY (Matrix) </math>, <math>Matrix</math> is set of values to find the factorization value.
−
*Partition matrices in as <math>A</math> = <math>LL^{T}</math>  
+
*Partition matrices in <math>A</math> = <math>LL^{T}</math> is
 +
<math>
 +
\begin{bmatrix}
 +
a_{11} & A_{21}^{T}\\
 +
A_{21} & A_{22}
 +
\end{bmatrix}
 +
=
 +
\begin{bmatrix}
 +
l_{11} & 0\\
 +
L_{21} & L_{22}
 +
\end{bmatrix}
  
−
==ZOS Section==
+
\begin{bmatrix}
 +
l_{11} & L_{21}^{T}\\
 +
0  & L_{22}^{T}
 +
\end{bmatrix}
 +
=
 +
\begin{bmatrix}
 +
l_{11}^{2} & L_{11}L_{21}^{T}\\
 +
L_{11}L_{21} & L_{21}L_{21}^{T} + L_{22}L_{22}^{T}
 +
\end{bmatrix}
 +
</math>
 +
 
 +
:<h2>Algorithm</h2>
 +
<ol>
 +
<li>Determine <math>l_{11}</math> and <math>L_{21}</math></li>
 +
<math>l_{11}</math> = <math>\sqrt{a_{11}}</math>
 +
<math>L_{21}</math> = <math>\frac{1}{l_{11}}A_{21}</math>
 +
<li>Compute <math>L_{22}</math> from </li>
 +
<math>A_{22}-L_{21}L_{21}^{T}</math> = <math>L_{22}L_{22}^{T}</math>
 +
*This is a Cholesky Factorization of order <math>n-1</math>
 +
</ol>
  
 
==Examples==
 
==Examples==
−
CHOLESKY([[16,32,12],[12, 18, 0],[ -5, 0, 11]])  
+
{| class="wikitable"
 +
|+Spreadsheet
 +
|-
 +
! !! A !! B !! C     
 +
|-
 +
! 1
 +
| 16 || 32 || 12
 +
|-
 +
! 2
 +
| 12 || 18 || 0
 +
|-
 +
! 3
 +
| -5 || 0 || 11
 +
|}
 +
=CHOLESKY(A1:C3)  
  
 
{| class="wikitable"
 
{| class="wikitable"
−
 
 
|+Result
 
|+Result
 
|-
 
|-
Line 30: Line 71:
 
|}
 
|}
  
−
CHOLESKY([[25, 15, -5],[15, 18, 0],[ -5, 0, 11]])
+
{| class="wikitable"
 +
|+Spreadsheet
 +
|-
 +
! !! A !! B !! C     
 +
|-
 +
! 1
 +
| 25 || 15 || -5
 +
|-
 +
! 2
 +
| 15 || 18 || 0
 +
|-
 +
! 3
 +
| -5 || 0 || 11
 +
|}
 +
=CHOLESKY(A1:C3)
  
 
{| class="wikitable"
 
{| class="wikitable"
Line 41: Line 96:
 
| -1 || 1 || 3
 
| -1 || 1 || 3
 
|}
 
|}
 +
 +
==Related Videos==
 +
 +
{{#ev:youtube|v=gFaOa4M12KU|280|center|Cholesky Decomposition}}
  
 
==See Also==
 
==See Also==
 +
*[[Manuals/calci/ANTIDIAGONAL| ANTIDIAGONAL]]
 +
*[[Manuals/calci/CONFERENCE| CONFERENCE]]
 +
*[[Manuals/calci/PASCAL| PASCAL]]
 +
 +
==References==
 +
*[http://www.seas.ucla.edu/~vandenbe/103/lectures/chol.pdf Cholesky Factorization]
 +
 +
 +
 +
*[[Z_API_Functions | List of Main Z Functions]]
 +
 +
*[[ Z3 |  Z3 home ]]

Latest revision as of 14:55, 26 November 2018

CHOLESKY (Matrix)


  • is the array of numeric elements.

Description

  • This function gives the value of Cholesky factorization.
  • It is called Cholesky Decomposition or Cholesky Factorization.
  • The Cholesky Factorization is only defined for symmetric or Hermitian positive definite matrices.
  • Every positive definite matrix A can be factored as = 
 is lower triangular with positive diagonal elements
 is is the conjugate transpose value of
  • Every Hermitian positive-definite matrix has a unique Cholesky decomposition.
  • Here , is set of values to find the factorization value.
  • Partition matrices in = is

Algorithm

  1. Determine and
    2. = =
  2. Compute from
    3. =
    • This is a Cholesky Factorization of order

Examples

Spreadsheet
A B C
1 16 32 12
2 12 18 0
3 -5 0 11

=CHOLESKY(A1:C3)

Result
4 0 0
3 3 0
-1.25 1.25 2.80624
Spreadsheet
A B C
1 25 15 -5
2 15 18 0
3 -5 0 11

=CHOLESKY(A1:C3)

Result
5 0 0
3 3 0
-1 1 3

Related Videos

Cholesky Decomposition

See Also

References


Retrieved from "http://wiki.zcubes.com/index.php?title=Manuals/calci/CHOLESKY&oldid=211380"