Solve Endomorphism Sequence Questions in a Vectoriel Space

  • Context: Graduate 
  • Thread starter Thread starter Calabi
  • Start date Start date
  • Tags Tags
    Sequence
Click For Summary

Discussion Overview

The discussion revolves around solving problems related to endomorphisms in a finite normed vector space. Participants explore properties of linear operators, specifically focusing on kernels, images, and convergence of sequences of operators. The scope includes theoretical aspects and mathematical reasoning.

Discussion Character

  • Technical explanation
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • Some participants propose that for a linear operator \( u \) on a finite normed vector space, \( \text{Ker}(u - \text{Id}) = \text{Ker}((u - \text{Id})^{2}) \) holds, but they express uncertainty about proving the reverse inclusion.
  • There is a suggestion that if \( x \in \text{Ker}((u - \text{Id})^{2}) \) but not in \( \text{Ker}(u - \text{Id}) \), then \( y = (u - \text{Id})(x) \) is an eigenvector of \( u \) with eigenvalue 1, leading to discussions about linear independence and orthonormality of vectors.
  • Participants discuss the convergence of the sequence \( u_n = \frac{1}{n}(\text{Id} + \sum_{i=2}^{n} u^{i-1}) \) and its implications for projecting onto \( \text{Ker}(u - \text{Id}) \), with some expressing confusion about the nature of this convergence.
  • There are attempts to establish bounds on the norms of \( u_n(y) \) for \( y \in \text{Im}(u - \text{Id}) \) and to relate this to the convergence of \( u_n \) to a projection operator.
  • Some participants question the implications of certain results and seek clarification on the relationships between different mathematical expressions and concepts discussed.

Areas of Agreement / Disagreement

Participants generally agree on the properties of the kernel and image of the operator \( u \), but there are multiple competing views on the implications of these properties, particularly regarding convergence and projection. The discussion remains unresolved on several technical points.

Contextual Notes

Some participants express uncertainty about the definitions and implications of orthonormality in the context of the vectors discussed. There are also unresolved mathematical steps related to the convergence of sequences and the properties of linear operators.

Calabi
Messages
140
Reaction score
2
Hello let be a finish normed vectoriel space $$(E, ||.||)$$, and $$u \in L(E) / ||u|| \leq 1$$.
1) Show that $$Ker(u - Id) = Ker((u - Id)^{2})$$.
For $$\subset$$ it's obvious, but for $$\supset$$ I don't know. I suppose $$\exists x_{1} \in Ker((u - Id)^{2})\Ker(u - Id)$$. So I can say that $$u(x_{1}) - x_{1} \neq 0_{E}$$ But I don't know.

2) Show that $$E = Im(u - Id) \oplus Ker(u - Id)$$
If I have an $$x \in Im(u - Id) \cap Ker(u - Id)$$, I have a w in E and I have $$u(w) - w = x$$ so $$u(x) - x = 0_{E} = (u - Id)^{2}(w)$$, so $$w \in Ker((u - Id)^{2}) = Ker(u - Id)$$ so $$x = 0_{E}$$.
And like I have $$dim(E) = rg(u - Id) + dim(Ker(u - Id))$$. It conclude.

3) Show that $$u_{n} = \frac{1}{n}(Id + \sum_{i=2}^{n} u^{i-1})$$ converge in an projector on $$Ker(u - Id)$$.
I don't know. By the previous question I can say that if I fix an x $$Ker(u - Id)$$ like u(x) = x,
$$u_{n}(x) = x$$. But like with the image I find nothing I give up with this idea.

Could you help me please?

Thank you in advance and have a nice afternoon:oldbiggrin:.
 
Physics news on Phys.org
Calabi said:
Hello let be a finish normed vectoriel space ##(E, ||.||)##, and ##u \in L(E) / ||u|| \leq 1##.
1) Show that ##Ker(u - Id) = Ker((u - Id)^{2})##.
For ##\subset## it's obvious, but for ##\supset## I don't know. I suppose ##\exists x_{1} \in Ker((u - Id)^{2})\Ker(u - Id)##. So I can say that ##u(x_{1}) - x_{1} \neq 0_{E}## But I don't know.

Let's start with this. Take ##x\in \text{Ker}((u-\text{Id})^2)\setminus \text{Ker}(u-\text{Id})##.
1) Show that ##y=(u -\text{Id})(x)## is an eigenvector of ##u## with eigenvalue ##1##.
2) Show that ##\{x,y\}## are linearly independent. Show that we can find ##x## and ##y## to be orthonormal.
3) Calculate ##u(x)## in terms of the "basis" ##\{x,y\}##.
4) Find a lower bound for ##\|u(x)\|## and use it to show that ##\|u\|>1##.
 
Calabi said:
3) Show that ##u_{n} = \frac{1}{n}(Id + \sum_{i=2}^{n} u^{i-1})## converge in an projector on ##Ker(u - Id)##.
I don't know. By the previous question I can say that if I fix an x ##Ker(u - Id)## like u(x) = x,
##u_{n}(x) = x##. But like with the image I find nothing I give up with this idea.

For this, consider an element ##y= u(w) - w## in the image of ##u - \text{Id}##. Can you rewrite

[tex]y + u(y) + ... + u^{n-1}(y)[/tex]

in a suitable way? Think of telescoping sums.
 
Hello and thanks for answer, for 1) : $$y = u(x) - x$$ so $$u(y) = y$$ because $$x \in \text{Ker}((u-\text{Id})^2)$$.
Then $$ax + by = 0_{E} \Rightarrow a(u(x) - x) = 0_{E} \Rightarrow a = 0$$ because $$u(x) - x \neq 0_{E}$$ as $$
x\in \text{Ker}((u-\text{Id})^2)\setminus \text{Ker}(u-\text{Id})(1)$$. And because of (1) $$y \neq 0_{E}$$ so $$b = 0$$. I don't see what you means by orthonormal. So in the subspace geberae by x and y, the matrix of $$u - Id$$ is $$\beginn{pmatrix}(0 1\\11 )\end{pmatrix}$$. For 4) I don't know. Could you help me more pkease?

For 3) as $$y \in Im(u - Id) $$ as you say we get $$u_{n}(y) = \frac{1}{n}(u^{n}(w) - w)$$. Coukld you help me more please?

Thank you in advance and have a nice morning:oldbiggrin:.
 
So if I fix a $$y \in Im(u - Id)$$, I have $$\forall n \in \mathbb{N}, ||u_{n}(y)|| \leq \frac{1}{n}||u^{n}(w)|| + \frac{1}{n}||w|| = \frac{2}{n}||w||$$ because $$||u|| \leq 1$$ and $$||u^{n}|| \leq ||u||^{n}$$. So lim $$u_{n}(y) = 0_{E}$$.
So for $$x \in E, x = a + b, a \in Ker(u - Id), b \in Im(u - Id), lim u_{n}(x) = a$$. But why does a such simple convergence make that my sequence of $$u_{n}$$ converge****** in a projector on $$Ker(u - Id)$$ please? ****** For the norme $$||u|| = sup_{x \neq 0_{E}} \frac{||u(x)||}{||x||}$$ in $$L(E)$$.

Thank you in advance and have a nice afternoon:oldbiggrin:.
 
Calabi said:
Hello and thanks for answer, for 1) : $$y = u(x) - x$$ so $$u(y) = y$$ because $$x \in \text{Ker}((u-\text{Id})^2)$$.
Then $$ax + by = 0_{E} \Rightarrow a(u(x) - x) = 0_{E} \Rightarrow a = 0$$

What happened to ##b##?

I don't see what you means by orthonormal. So in the subspace geberae by x and y, the matrix of $$u - Id$$ is $$\beginn{pmatrix}(0 1\\11 )\end{pmatrix}$$.

For 4) I don't know. Could you help me more pkease?

Try to prove that ##\|Ay\|>\|y\|##.
 
Calabi said:
So if I fix a $$y \in Im(u - Id)$$, I have $$\forall n \in \mathbb{N}, ||u_{n}(y)|| \leq \frac{1}{n}||u^{n}(w)|| + \frac{1}{n}||w|| = \frac{2}{n}||w||$$ because $$||u|| \leq 1$$ and $$||u^{n}|| \leq ||u||^{n}$$. So lim $$u_{n}(y) = 0_{E}$$.
So for $$x \in E, x = a + b, a \in Ker(u - Id), b \in Im(u - Id), lim u_{n}(x) = a$$. But why does a such simple convergence make that my sequence of $$u_{n}$$ converge****** in a projector on $$Ker(u - Id)$$ please?

So you see that the pointswise limit ##u(x) = \lim_n u_n(x)## is a projection. Try to find a simple expression for ##u(x) - u_n(x)## and use this to show that ##\|u - u_n\|\rightarrow 0##.
 
micromass said:
What happened to bb?
I said after we have $$by = 0_{E}$$ and y is not nul otherwise $$x \in Ker(u - Id)$$. SO b is nul.
micromass said:
So you see that the pointswise limit u(x)=limnun(x)u(x) = \lim_n u_n(x) is a projection. Try to find a simple expression for u(x)−un(x)u(x) - u_n(x) and use this to show that ∥uun∥→0\|u - u_n\|\rightarrow 0.

OK but it's just said me that for a certain x(fixed.). the sequence $$u_{n}(x)$$ have the projection of x on $$Ker(u - Id)$$ I can call it $$p(x)$$ so p(x) define an application which is linear. I really don't see the link with $$||u - u _{n}|| =
sup_{x \neq 0_{E}} \frac{||u(x) - u_{n}(x)||}{||x||}$$. Could you help me please?

Oh by the way how to stop latex from go at the line each time I rigght somethnig please?Thank you in advance and have a nice afternoon:oldbiggrin:.
 

Similar threads

  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 2 ·
Replies
2
Views
3K
Replies
2
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 11 ·
Replies
11
Views
3K
  • · Replies 8 ·
Replies
8
Views
3K
  • · Replies 15 ·
Replies
15
Views
2K