Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
158 views

Quadratic Form Example

1. To compute arbitrary minors of order m (1 ≤ m ≤ n) for an n×n matrix A: fix m elements on the principal diagonal, delete rows and columns without fixed elements, and take the determinant of the resulting submatrix. 2. To check if A is positive/negative semidefinite, one must compute all arbitrary minors up to order n and check if they satisfy the necessary conditions. 3. For the example matrix A, some minors are negative, so A is neither positive nor negative semidefinite. The quadratic form associated with A is indefinite.

Uploaded by

api-3737025
Copyright
© Attribution Non-Commercial (BY-NC)
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
158 views

Quadratic Form Example

1. To compute arbitrary minors of order m (1 ≤ m ≤ n) for an n×n matrix A: fix m elements on the principal diagonal, delete rows and columns without fixed elements, and take the determinant of the resulting submatrix. 2. To check if A is positive/negative semidefinite, one must compute all arbitrary minors up to order n and check if they satisfy the necessary conditions. 3. For the example matrix A, some minors are negative, so A is neither positive nor negative semidefinite. The quadratic form associated with A is indefinite.

Uploaded by

api-3737025
Copyright
© Attribution Non-Commercial (BY-NC)
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

General rule to compute arbitrary minors:

2 3
a11 a21 ::: an1
6 a12 a22 7
Take matrix A = 6 4 :
7
5
:::
a1n ann
2 3
a11
6 a22 7
Consider principle diagonal of matrix A : 6
4
7
5
:::
ann

To compute arbitrary minors of order m (1 m n) :


1. Fix m elements of principle diagonal.
2. Delete all columns and all rows of matrix A that do not contain elements
of principle diagonal that you have …xed. You are left with submatrix which
have your chosen elements on principle diagonal.
3. Compute determinant of this submatrix.
n!
Note: there are m!(n m)! arbitrary minors of order m;where k! = 1 2 ::: k:
4. To show that matrix A positive(negative) semide…nite you have to check
all arbitrary minors for each m: 1 m n: To show that matrix A is not
positive(negative) semide…nite you have to …nd one arbitrary minor that does
not satisfy condition m 0 (( 1)m m 0).

Example 1 Consider quadratic form

Q = x21 + x22 + 4x23 2x1 x2 4x1 x3 + 10x2 x3


1. Write down this quadratic form in matrix notation
2 30 1
1 1 2 x1
Q = x0 Ax = x1 x2 x3 4 1 1 5 5 @ x2 A
2 5 4 x3

2. Find all principal minors of matrix A: Is matrix A positive/negative

de…nite?
D1 = 1

1 1
D2 = =1 1 ( 1) ( 1) = 0
1 1

1 1 2
1 5 1 2
D3 = 1 1 5 = ( 1)1+1 1 +( 1)2+1 ( 1) +
5 4 5 4
2 5 4
1 2
( 1)3+1 ( 2) = 21 + 6 + 6 = 9
1 5

1
D1 > 0; D2 = 0; D3 < 0
Matrix is neither positive de…nite nor negative de…nite
3. Find all arbitrary minors of matrix A: Is matrix A positive/negative
semide…nite?
m=1
Fix a11 : 1 =1>0
Fix a22 : 1 =1>0
Fix a33 : 1 =4>0

m=2
1 1
Fix a11 ; a22 : 2 = =0
1 1

1 2
Fix a11 ; a33 : 2 = =0
2 4

1 5
Fix a22 ; a33 : 2 = = 21 < 0
5 4

m=3
1 2 1
Fix a11 ; a22 ; a33 : 1 13 = 5 = 9<0
2 5 4
Matrix is neither positive semide…nite nor negative semide…nite.

Quadratic form is not PD, ND, PSD, NSD, hence, quadratic form is indef-
inite

Note: In this case to show that matrix is neither positive semide…nite nor
negative semide…nite it is not necessary to compute all arbitrary minors. We
know that 3 = det jAj = D3 : Before we found that D3 < 0: =)matrix A can
not be positive semide…nite (necessary condition is m 0; 80 m n). As
long as 1 > 0 matrix A can not be negative semide…nite (necessary condition
is ( 1)m m 0; 80 m n; as long as m = 1 ( 1)1 1 = 1 0).

You might also like