• Support PF! Buy your school textbooks, materials and every day products via PF Here!

Inner Prod Space, Positive Definite Proof

1. Homework Statement

Let H be an inner product space.

Let T:H->H be a linear, self adjoint, positive definite operator.

Fix h in H and let g = T(h) / square root (1 + (T(h),h)). for h in H

Show that the operator S:H->H defined by S(v) = T(v) - (v,g)g for v in H is positive definite (and self adjoint, but I already proved that part).
2. Homework Equations

L is self adjoint if L = L*
and by proposition FOR ANY L, (L(v),u) = (v,L*(u))
but => (L(v),u) = (v,L(u)) for self adjoint operators.

L is a positive definite operator <=> (L(v),v) > 0 for all v in V (where V is an inner product space).

3. The Attempt at a Solution

So far I've tried to show (S(v),v) > 0 by substituting in T(v) - (v,g)g.

When I do that I get:

(S(v),v) = (T(v) - (v,g)g,v) =

(T(v),v) - ((v,g)g,v)= (T(v),v) - (v,g)*(g,v).

Then I substitute in for g:

(T(v),v) - (v,T(h)/square root(1 + (T(h),h))*(T(h)/square root(1 + T(h),h)),v)

And I realize (since T is positive definite) that the above expression is greater than:

(T(v),v) - (v,T(h))*(T(h),v).
since square root 1 + a positive number > 1, and pulling the 1/square root(1+(T(h),h)) out of the two inner products.

I don't see this as helping all that much though and I am now stuck.
 

Want to reply to this thread?

"Inner Prod Space, Positive Definite Proof" You must log in or register to reply here.

Related Threads for: Inner Prod Space, Positive Definite Proof

Replies
5
Views
582
  • Posted
Replies
15
Views
4K
  • Posted
Replies
4
Views
807
  • Posted
Replies
18
Views
6K
  • Posted
Replies
2
Views
1K

Physics Forums Values

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving
Top