All the ways to build positive definite matrices

Click For Summary
SUMMARY

This discussion outlines multiple methods for constructing positive definite matrices, specifically square real matrices. Key methods include generating matrices from vectors using the formula vTivj, forming matrices via MMT, and utilizing diagonal matrices with positive entries. Additional methods include the sum and product of positive definite matrices, as well as the formation of block matrices and principal minors. The discussion emphasizes the versatility of positive definite matrices in various mathematical contexts.

PREREQUISITES
  • Understanding of matrix operations, including transpose and multiplication
  • Familiarity with diagonal and diagonal dominant matrices
  • Knowledge of positive definite matrix properties
  • Basic concepts of linear algebra and quadratic forms
NEXT STEPS
  • Research the properties of positive definite matrices in linear algebra
  • Explore the Cholesky decomposition for generating positive definite matrices
  • Learn about the implications of positive definiteness in optimization problems
  • Investigate the role of eigenvalues in determining matrix definiteness
USEFUL FOR

Mathematicians, data scientists, and engineers working with linear algebra, optimization, and statistical modeling will benefit from this discussion on positive definite matrices.

lukluk
Messages
8
Reaction score
0
Often people asks how to obtain a positive definite matrix. I would like to make a list of all possible ways to generate positive definite matrices (I consider only square real matrices here). Please help me to complete it.

Here M is any matrix, P any positive definite matrix and D any diagonal matrix.

1) given a vector v, build a matrix with entries
vTivj
(T is transpose)
2) given any matrix M, build
MMT
3) any diagonal matrix with positive entries
4) any diagonal dominant matrix with positive entries
5) the product
M D MT
where D is a diagonal matrix with positive entries
6) if P1,P2 are positive definite matrices then
P1 P2
P1 +P2
are also positive definite
7) if P is positive definite then also
M P M-1
P-1
are positive definite
9) the matrix with blocks
P1 M
MT P2
is positive definite
10) the product
aP
of a positive scalar a times a positive definite matrix P
11) any submatrix formed as the upper left square matrix of a positive definite matrix (principal minor) is also
a positive definite matrix

...

do you know other ways not trivially reconducible to one of the above?
 
Last edited:
Physics news on Phys.org
You want to solve the quadratic polynomial equation ##\displaystyle{\sum_{i,j}} a_{ij}x_ix_j>0## for all possible ##x_\, , \,\text{not all }x_i=0##. If we let run the ##x## through ##(1,0,\ldots,0),\ldots,(0,\ldots,0,1)## then we get ##n## linear constraints for ##n^2## variables ##a_{ij}##. These are really many possible solutions. E.g. any symmetric real square matrix ##A## is positive definite, if ##A=GG^\tau ## with a regular lower triangular matrix. See also https://en.wikipedia.org/wiki/Definiteness_of_a_matrix
 

Similar threads

  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 13 ·
Replies
13
Views
4K
  • · Replies 33 ·
2
Replies
33
Views
3K
  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 8 ·
Replies
8
Views
3K
  • · Replies 4 ·
Replies
4
Views
3K