Linear application and différential

Click For Summary

Homework Help Overview

The discussion revolves around a problem in functional analysis involving a normed vector space E with a specified dimension m. The original poster is tasked with demonstrating a relationship between the norms of a linear operator u and the identity operator Id, particularly focusing on the implications of the dimension of the space.

Discussion Character

  • Exploratory, Conceptual clarification, Assumption checking

Approaches and Questions Raised

  • The original poster attempts to construct a C^1 application with a continuous differential but expresses confusion about the relevance of the dimension m. Some participants suggest considering the eigenvalues of u, noting that there can be at most m eigenvalues, while others question the applicability of this approach in the context of the problem.

Discussion Status

Participants are exploring various ideas, including the use of the inverse function theorem and the implications of the dimension of the space for the uniqueness of solutions. There is a recognition that the initial approach may not be viable, and some participants are reconsidering their reasoning regarding the role of m.

Contextual Notes

There is a mention of potential constraints related to the properties of matrices in R, such as the possibility of lacking eigenvalues and the non-diagonalizability of certain matrices. The problem is situated within a set of differential calculus problems, which may impose additional considerations.

Calabi
Messages
140
Reaction score
2

Homework Statement


Let be E a normded vectoriel space, ##dim(E) = m \in \mathbb{N}^{*}##, I have to show that ##\exists \rho_{1}, \rho_{2} > 0 | \forall u \in L(E), ||u^{m} - Id| \leq \rho_{1} \Rightarrow |u - Id| \leq \rho_{2}##.

Homework Equations


Nothing.

The Attempt at a Solution


[/B]
I try to constrcut a ##C^{1}## application which will have by definition a continuous différential. But I don't see what the dimension m as to do in here and I'm really locked.

Could you give me a clue please?

Thank you in advance and have a nice afternoon:oldbiggrin:.
 
Physics news on Phys.org
Think about the eigenvalues of [itex]u[/itex]. There are at most [itex]m[/itex] of them, and a finite set of real numbers always has a maximum and a minimum.
 
If I look in ##\mathbb{R}## it's possible ther's no eigenvalues and the matrice I ut is not necessarely diagonaloisabled.
Are you sure it's really exploitable?

I find this problem on differential calcul set of problem.
 
I have an idea.
 
Let put us around the identity ##Id##, the application ##\phi : u \in L(E) \rightarrow u^{m} \in L(E)## is ##C^{\infty}##, as the differential in ##Id## is an isomorphism, then the inverse function theorem show that ##\exists \rho_{1}, \rho_{2} > 0 | \phi : B(Id, \rho_{1}) \rightarrow B(Id, \rho_{2})## is a bijection. But as ##\phi## is not injective it doesn't work.
 
I corrrecte the ennoncee : ##
\exists \rho_{1}, \rho_{2} > 0 | \forall v \in L(E) \exists u \in L(E) | u^{m} = v \text{ and } ||v - Id|| \leq \rho_{1} \Rightarrow ||u - Id|| \leq \rho_{2}##
 
I think we use m is the dimension for the unicity but I don't see why, the rest is an application of the inverse function theorem.
Thanks for your help.
 
Yeah what I wroght non works I think but I don't use ##m = dim(E)##.
What do you think please?
 

Similar threads

  • · Replies 7 ·
Replies
7
Views
1K
Replies
15
Views
3K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
Replies
1
Views
2K
  • · Replies 10 ·
Replies
10
Views
4K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 12 ·
Replies
12
Views
2K
  • · Replies 11 ·
Replies
11
Views
2K
Replies
9
Views
2K