Positive definite matrix bounded below

Click For Summary

Homework Help Overview

The discussion revolves around the properties of a positive definite matrix and its relation to a quadratic polynomial defined as Q(x)=\frac{1}{2}\langle{x, Ax\rangle}-\langle{b, x\rangle}. Participants are tasked with demonstrating that this polynomial is bounded below for all x in R^n.

Discussion Character

  • Exploratory, Conceptual clarification, Mathematical reasoning

Approaches and Questions Raised

  • The original poster attempts to find a constant m such that Q(x) is greater than or equal to m for all x. They draw parallels to the properties of parabolas but express uncertainty in generalizing this to quadratic forms. Other participants introduce the concept of a matrix C such that C^2=A and question the existence and invertibility of C. They also explore the implications of these definitions on the evaluation of the quadratic form.

Discussion Status

Participants are actively engaging with the problem, with some offering insights into the properties of positive definite matrices and their implications for the quadratic form. There is a productive exchange of ideas, particularly regarding the evaluation of expressions involving the matrix C and its relationship to the original problem.

Contextual Notes

There are ongoing questions about the existence of the matrix C and its invertibility, which are tied to the properties of the positive definite matrix A. These aspects remain under discussion without resolution.

Charles49
Messages
87
Reaction score
0

Homework Statement



Let [tex]A[/tex] be a positive definite [tex]n\times n[/tex] real matrix, [tex]b\in\mathbb{R}^n[/tex], and consider the quadratic polynomial [tex]Q(x)=\frac{1}{2}\langle{x, Ax\rangle}-\langle{b, x\rangle}[/tex]. Show that [tex]Q[/tex] is bounded below.

2. The attempt at a solution

I have to come up with a constant [tex]m[/tex] so that [tex]Q(x)\ge m[/tex] for all [tex]x\in\mathbb{R}^n.[/tex] I see that [tex]Q[/tex] looks a lot like a parabola. I know how to find the lower bound of a parabola opening upward but I don't know how to generalize this to quadratic forms.
 
Last edited:
Physics news on Phys.org
What is a??
 
Sorry, its supposed to be x.
 
Let C be such that [itex]C^2=A[/itex] (why does C exist?) and let [itex]c=C^{-1}b[/itex] (why is C invertible?).

Calculate

[tex]\frac{1}{2}<Cx-c,Cx-c>[/tex]
 
This is what I got but I don't know what it does

[tex]\frac{1}{2}\langle{Cx-C^{-1}b, Cx-C^{-1}b\rangle} = \frac{1}{2}||Cx-C^{-1}b||\\<br /> =\frac{1}{2}||C^{-1}(Ax-b)||[/tex]

[tex]C[/tex] exists because A is positive definite. [tex]C^{-1}[\tex] exists because A has no zero eigenvalues.[/tex]
 
Last edited:
Yes, so that shows that

[tex]<Cx-C^{-1}b,Cx-X^{-1}b>\geq 0[/tex]

Now try to evaluate this in another way... Use the linearity of the inproduct.
 
This suggestion helped a lot. Thanks.
 

Similar threads

  • · Replies 4 ·
Replies
4
Views
1K
Replies
7
Views
2K
Replies
20
Views
4K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 43 ·
2
Replies
43
Views
5K