Question on decomposition of a matrix

  • Context: Graduate 
  • Thread starter Thread starter mnb96
  • Start date Start date
  • Tags Tags
    Decomposition Matrix
Click For Summary
SUMMARY

The discussion centers on the decomposition of a 2x2 symmetric positive definite matrix M, expressed as M = A^T Σ A, where Σ is also symmetric positive definite and A is a nonsingular matrix. The goal is to find a matrix P_M such that P_M = AQ, with Q independent of A. The Cholesky decomposition is referenced, specifically M = LDL^τ, where D is diagonal and L is lower triangular. The transformation S = LD^{1/2} allows for the representation M = SS^τ, providing a pathway to derive conditions for 2x2 matrices.

PREREQUISITES
  • Understanding of matrix decomposition techniques, specifically Cholesky decomposition.
  • Knowledge of properties of symmetric positive definite matrices.
  • Familiarity with matrix multiplication and transformations.
  • Basic linear algebra concepts, particularly related to nonsingular matrices.
NEXT STEPS
  • Study the Cholesky decomposition in detail, focusing on its applications to symmetric positive definite matrices.
  • Explore the properties and applications of symmetric positive definite matrices in various mathematical contexts.
  • Research matrix transformations and their implications in linear algebra.
  • Investigate the implications of nonsingular matrices in matrix equations and decompositions.
USEFUL FOR

Mathematicians, data scientists, and engineers working with linear algebra, particularly those involved in matrix analysis and decomposition techniques.

mnb96
Messages
711
Reaction score
5
Hello,

I have a [itex]2\times 2[/itex] real matrix [itex]M[/itex] such that: [tex]M=A^T \Sigma A[/tex], where the matrix [itex]\Sigma[/itex] is symmetric positive definite, and [itex]A[/itex] is an arbitrary 2x2 nonsingular matrix. Both A and ∑ are unknown, and I only know the entries of the matrix M itself. Note that M is symmetric positive definite too.

I was wondering if it is possible to apply some decomposition of the matrix [itex]M[/itex] in order to find another matrix [itex]P_M[/itex] such that: [tex]P_M = AQ[/tex]
where the matrix Q must not depend on A (e.g. it cannot be a product of matrices where A appears). I basically want to find a matrix PM where the multiplication with A appears only at the left side, and not at both sides like in M.
 
Last edited:
Physics news on Phys.org
The Cholesky normal form for symmetric positive definite matrices ##M## gives us ##M=LDL^\tau## as yours. But we have additionally that ##D## is diagonal and ##L## a lower triangular matrix. If we take ##D^{1/2}\cdot D^{1/2}=D## the square root of ##D## and set ##S:=LD^{1/2}## we can even write ##M=SS^\tau## with a lower triangular matrix ##S##. For ##2\times 2## matrices, this should give you some nice conditions.
 

Similar threads

  • · Replies 33 ·
2
Replies
33
Views
3K
  • · Replies 13 ·
Replies
13
Views
4K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 13 ·
Replies
13
Views
4K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 8 ·
Replies
8
Views
3K