Scribd
Upload
348 views2 pages

Positive Definite Matrix Tests

There are three main ways to determine if a real symmetric matrix A is positive semidefinite (PSD) or positive definite (PD): 1) use the definition and check if y'Ay is nonnegative/positive for all vectors y, 2) use eigenvalues - A is PSD if all eigenvalues are nonnegative and PD if all are positive, 3) use determinants - A is PSD if all principal minors are nonnegative and PD if all leading principal minors are positive. The document provides examples to illustrate these tests.

Uploaded by

fatihy73
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
348 views2 pages

Positive Definite Matrix Tests

There are three main ways to determine if a real symmetric matrix A is positive semidefinite (PSD) or positive definite (PD): 1) use the definition and check if y'Ay is nonnegative/positive for all vectors y, 2) use eigenvalues - A is PSD if all eigenvalues are nonnegative and PD if all are positive, 3) use determinants - A is PSD if all principal minors are nonnegative and PD if all leading principal minors are positive. The document provides examples to illustrate these tests.

Uploaded by

fatihy73
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Positive (Semi)Definite Matrices

There are at least three ways of determining whether a real symmetric matrix A is positive
semidefinite (PSD) or positive definite (PD).

1. Use the definition:


y ' Ay  0 y  R n (PSD)
y ' Ay  0 y  R n :y  0 (PD)

This can be difficult since you have to show that y 'Ay is nonnegative or positive for all vectors y.

2. Use eigenvalues:
If all eigenvalues of A are strictly positive, then A is positive definite.
If all eigenvalues of A are nonnegative, then A is positive semidefinite.
This method can be easy if you have software that computes eigenvalues.

3. Use determinants:
A is positive semidefinite if and only if all of its principal minors are nonnegative. Recall that a
principal minor is the determinant of a submatrix formed by removing an equal number of rows
and columns; see http://en.wikipedia.org/wiki/Principal_minor . The rows removed must
correspond to the columns removed; for example, if rows 2 and 4 are removed, columns 2 and
4 must also be removed.

A is positive definite if and only if all of its leading principal minors are positive. Recall that a
leading principal minor is the determinant of a submatrix formed by removing an equal number
of rows and columns “from the back.” That is, a leading principal minor is the determinant of an
upper-left submatrix.

Example: Consider the 3-by-3 matrix


1 2 3
4 5 6
 
 7 8 9 

This matrix has the following principal minors:

1 2 3
 4 5 6  ,  1 2  ,  1 3  , 5 6  , 1 , 5 , 9
   4 5   7 9  8 9       
7 8 9       

And the following leading principal minors:

1 2 3
 4 5 6 , 1 2  , 1
  4 5  
7 8 9   
The determinant test is easy to do on small homework/exam problems with 2-by-2 matrices.
A 2-by-2 symmetric matrix A is positive semidefinite if both diagonal terms are nonnegative and
the determinant is nonnegative. It is positive definite if the upper-left element is positive and the
determinant is positive.

A note on all methods: if you’re asked to check whether the Hessian  2 f ( x) is positive
(semi)definite, you may need to check whether it is positive (semi)definite for all x.

Here are some practice matrices:

2 2
2 4 (positive definite)
 

 1 2 
 2 4  (positive semidefinite)
 


 1 3  (not positive semidefinite)
 3 2 
  




1 1 1  
1 1 1  (not positive semidefinite)
  

1 1 0  




 x 1 1 
 1 1 1  (not positive semidefinite for all x)
 
 1 1 0 

You might also like