Homeomorphism in a Banach space

Click For Summary

Homework Help Overview

The discussion revolves around a problem in functional analysis concerning a Banach space, specifically focusing on the properties of a continuous automorphism and a Lipschitz function. The original poster seeks to demonstrate that the sum of the automorphism and the Lipschitz function is bijective and that its inverse is continuous.

Discussion Character

  • Exploratory, Conceptual clarification, Mathematical reasoning

Approaches and Questions Raised

  • Participants discuss the application of the Banach fixed-point theorem as a potential approach to the problem. There are attempts to clarify the conditions under which the theorem can be applied, particularly regarding injectivity and surjectivity of the mappings involved.

Discussion Status

The discussion is ongoing, with various participants providing insights and suggestions. Some participants express uncertainty about specific steps in their reasoning, while others propose alternative approaches or clarifications. There is no explicit consensus yet, but several productive lines of inquiry have been initiated.

Contextual Notes

Participants note potential issues with the assumptions regarding the Lipschitz constant and the existence of certain norms. There are also discussions about the implications of specific properties of the functions involved, such as continuity and the nature of the mappings being considered.

Calabi
Messages
140
Reaction score
2
Member warned about posting with no effort shown
Hello,

1. Homework Statement

Let be E a banach space, A a continuous automorphsim(by the banach theorem his invert is continus too.). and f a k lipshitzian fonction with $$k < \frac{1}{||A^{-1}||}$$.

Homework Equations


$$k < \frac{1}{||A^{-1}||}$$

The Attempt at a Solution



I have to show that A + f is bijectiv and is invert is continuous too.

I have no real clue for that in the moment.
Could you help me please?

Thank ou in avnace and have a nice afternoon:oldbiggrin:.
 
Last edited by a moderator:
Physics news on Phys.org
Thanks to @Krylov , here is a tip: Banach fixed-point theorem.
 
  • Like
Likes   Reactions: S.G. Janssens
Hello and good year. Could you be more precise please?
 
Calabi said:
Hello and good year. Could you be more precise please?
Good year.

This tip is really all you need to solve the problem. You could look up what the Banach fixed-point theorem exactly says, and then see how it applies to this problem.
 
  • Like
Likes   Reactions: S.G. Janssens
OK I'm going to think about it.
 
Hello and have a nice year : first : $$||A^{-1} o f|| < ||A^{-1}|| ||f|| < kA^{-1} < 1$$ sol let be h in the continious fonction forme E to E with $$h o (Id + A^{-1}f) = (Id + A^{-1}of) o h = Id$$. Since A linear we've got $$(A + f) o h = A \Leftrightarrow (A+f)oh o A^{-1} = Id$$. I don't know if that mean there's an inverse at right.
At least my invert is continous.
 
Calabi said:
Hello and have a nice year : first : $$||A^{-1} o f|| < ||A^{-1}|| ||f|| < kA^{-1} < 1$$ sol let be h in the continious fonction forme E to E with $$h o (Id + A^{-1}f) = (Id + A^{-1}of) o h = Id$$. Since A linear we've got $$(A + f) o h = A \Leftrightarrow (A+f)oh o A^{-1} = Id$$. I don't know if that mean there's an inverse at right.
At least my invert is continous.
I'm not sure I understand this: for instance; where does this h come from?

Maybe someone else understand what you did and can comment, in that case just disregard what follows.

The way I would approach this (using @Krylov 's tip, but maybe not in the way he meant it):
First consider the simple case where A=I (the identity) and Lip(f)=k<1, and prove that I-f is invertible and the inverse is Lipschitz.

Proving that I-f is injective is easy (and yields the Lipschitz constant for the inverse of I-f).
EDIT: see the following posts.

In order to prove that I-f is surjective, use the Banach fixed-point theorem.
Hint: take y∈E, and consider the mapping Y: E→E defined by Y(x)=y+f(x). Can we apply the Banach fixed-point theorem to Y? And if so, what does it say about the surjectivity of I-f?

Again, maybe your proof is correct, if so, just disregard this.
 
Last edited:
Samy's suggestion is just fine, but I don't think it's necessary to consider surjectivity and injectivity separately.
 
Krylov said:
Samy's suggestion is just fine, but I don't think it's necessary to consider surjectivity and injectivity separately.
Oh yes, you get injectivity from the uniqueness of the fixed point.
 
  • Like
Likes   Reactions: S.G. Janssens
  • #10
Hello every body the fact is I'm not sure of my first equality. It's not even sur that $$||f||$$ exist(exept if f is nul in 0.).
At least I try something(even if i think it could be true.). anyway.
Let suppose i - f is not injctive and let's consider x in y in E with $$x \neq y and x - f(x) = y - f(x)$$
We get y = x which is absurde. Let's take this by the beginning : we want to show that y as got an antecedant by I - f.
Let's create $$\phi(x) = y + f(x)$$. It's clearly k < 1 lipshitzian. So let's applic the banach theorem, we find x in E with $$x - f(x) = E$$ which proove the surjectivity.

Since it's proove let's we can wright $$A + f = A o (I - A^{-1}of)$$
and by the hypothesis we've got $$A^{-1} o f$$ $$k||A^{-1}|| < 1$$ lipshitzian.
So it's conclude.
 
  • #11
Oh ther's also to proove that the inverse of I - f lipshitzian
 
  • #12
Calabi said:
Hello every body the fact is I'm not sure of my first equality. It's not even sur that $$||f||$$ exist(exept if f is nul in 0.).
At least I try something(even if i think it could be true.). anyway.
Let suppose i - f is not injctive and let's consider x in y in E with $$x \neq y and x - f(x) = y - f(x)$$
We get y = x which is absurde.
You get ##x-f(x)=y-f(y)##, not ##x - f(x) = y - f(x)##, so this part of your proof is not correct (or at least incomplete).
Calabi said:
Let's take this by the beginning : we want to show that y as got an antecedant by I - f.
Let's create $$\phi(x) = y + f(x)$$. It's clearly k < 1 lipshitzian. So let's applic the banach theorem, we find x in E with $$x - f(x) = E$$ which proove the surjectivity.
You probably meant that the Banach fixed-point theorem gives you a ##x\in E## for which ##x-f(x)=y## (not ##=E## as you wrote). This part is correct.
Calabi said:
Since it's proove let's we can wright $$A + f = A o (I - A^{-1}of)$$
and by the hypothesis we've got $$A^{-1} o f$$ $$k||A^{-1}|| < 1$$ lipshitzian.
So it's conclude.
Yes, this way you get the general case from the particular case where A=I. You have the sign wrong in ##A + f = A o (I - A^{-1}of)##, but that is no big deal, as what you proved is equally true for ##-f##.

You still have two issues left:
1) Injectivity (though that is almost done with what you have, the hint has already been given previously).
2) Show that the inverse is Lipschitz too.
 
Last edited:
  • #13
As it came up elsewhere, I'd like to add to this thread the proof that the inverse of ##I-f## is a Lipschitz function (under the assumptions of post #7).
Take ##x, y \in E##. Then, since ##f## is Lipschitz with Lipschitz constant ##k \lt 1##, ##\|f(x)-f(y)\|\leq k\|x-y\|##
Using the triangle inequality, one gets:
##\|(I-f)(x)-(I-f)(y)\|=\|x-y-(f(x)-f(y))\| \geq \|x-y\|-\|f(x)-f(y)\| \geq \|x-y\| -k\|x-y\|=(1-k)\|x-y\|##.
This shows that the inverse of ##I-f## is a Lipschitz function with Lipschitz constant ##\leq \frac{1}{1-k}##.
 
Last edited:
  • #14
Samy_A said:
This shows that the inverse of ##I−f## is a Lipschitz function with Lipschitz constant ##≤\frac{1}{1−k}##
Maybe I'm just too blind to see, but I didn't get it.
We have ##||(I-f)(x)-(I-f)(y)|| \geq q ||x-y||## for a ##1>q>0##. How does this turn into an upper bound?
 
  • #15
fresh_42 said:
Maybe I'm just too blind to see, but I didn't get it.
We have ##||(I-f)(x)-(I-f)(y)|| \geq q ||x-y||## for a ##q>0##. How does this turn into an upper bound?
I should have added the following.

Set ##T=(I-f)^{-1}##.
Take ##x', y' \in E##. Set ##x=Tx', y=Ty'##
As shown above, we have ##\|(I-f)(x)-(I-f)(y)\| \geq (1-k)\|x-y\|##.
This give ##\|x'-y'\| \geq (1-k)\|Tx'-Ty'\|##, or ##\|Tx'-Ty'\| \leq \frac{1}{1-k}\|x'-y'\|##.
 
Last edited:
  • #16
Thanks. I've just forgotten the "inverse of" part before ##I-f##.
Edit: you should have written ##(I-f)^{-1}## for dummies like me :frown:
 
  • #17
fresh_42 said:
Thanks. I've just forgotten the "inverse of" part before ##I-f##.
Edit: you should have written ##(I-f)^{-1}## for dummies like me :frown:
I agree (not about the "dummies" :smile:).
Let's rewrite it in terms of ##T=(I-f)^{-1}##. (As it is shorter and clearer.)

For ##x \in E##, ##(I-f)Tx=x##, so that ##Tx=x+f(Tx)##

Take any ##x,y \in E##:
##\|Tx-Ty\|=\|x+f(Tx)-y-f(Ty)\|=\|x-y+f(Tx)-f(Ty)\| \leq \|x-y\|+\|f(Tx)-f(Ty)\| \\ \leq \|x-y\| + k\|Tx-Ty\|##
Or ##\|Tx-Ty\|-k\|Tx-Ty\| =(1-k) \|Tx-Ty\| \leq \|x-y\|##.
Hence ##\|Tx-Ty\| \leq \frac{1}{1-k}\|x-y\|##, proving that ##T=(I-f)^{-1}## is a Lipschitz function.
 

Similar threads

Replies
2
Views
1K
  • · Replies 12 ·
Replies
12
Views
2K
  • · Replies 12 ·
Replies
12
Views
2K
  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 21 ·
Replies
21
Views
3K
  • · Replies 18 ·
Replies
18
Views
2K
Replies
1
Views
1K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 5 ·
Replies
5
Views
2K