Hilbert Space: Closest point property

P3X-018
Messages
144
Reaction score
0

Homework Statement



The theorem about the closest point property says:
If A is a convex, closed subspace of a hilbert space H, then

\forall x \in H\,\, \exists y \in A:\,\,\,\, \| x-y\| = \inf_{a\in A}\|x-a\|

I have to show that it is enough to show this theorem for x = 0 only, by using the isometry T_{x_0}(x) = x_0 + x.

The Attempt at a Solution



So I would have to show that \exists y \in A such that \|y\| = \inf_{a\in A}\|a\|, that is if A contains an element with least length, than for any point x in H there is a point in A that is closest to x, than any other in A.
Then what?
Any hint is appreciated.
 
Physics news on Phys.org
If A is a subspace, the claim \exists y so that \|y\| = \textrm{inf}_{a\in A} \|a\|, is trivial, because such y is the origo. So I suppose the A was not supposed to be assumed to be a subspace. A being convex and closed should be enough. Then... A remains convex and closed in translations.
 
Yeah that's right. A isn't assumed to be a subspace, just a subset.
And you are saying that "A remains convex and closed in translations", how can this be used to show what I want. Because x -a need not lie in A, so I can't use the property with least length. Or how would I be able to use it?
 
I needed to stop and think myself... but let's see, I think I got this.

One hint that I must give you, is that draw pictures! These are kind of tasks where there is too many different things that you can start doing, and it's difficult to guess what you are supposed to start doing without first drawing some kind of picture.

A set

<br /> -x + A = \{-x+a\;|\; a\in A\}<br />

seems to can handy. Choose y&#039;\in -x+A so that it has the smallest norm in this set. Then the desired y should be almost found.
 
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...

Similar threads

Back
Top