Hamburg University of Technology / Institute of Mathematics / 18th Internet Seminar / Discussion Board Lecture 06

Discussion Board Lecture 06

Remark 9.15

[#41]

Dear all,

In remark 9.15, I think you need, more than the continuos embedding, density of V in H for justifying why the weak convergence in V imply weak convergence in H. In fact, in section 9.4 there where we apply this remark we have this density (a is j-elliptic and j have dense range).

Best wishes, Abdallah.

Posted by Abdallahmaichine on 22 January 2015 at 14:36.

Sorry, this is for lecture 9.

Abdallah Maichine.

Posted by Abdallahmaichine on 22 January 2015 at 14:38.

Question on Example 6.16

[#37]

First, thank you for these courses. I need the proof of the fact that V is a Hilbert space?

Posted by Abdallahmaichine on 19 January 2015 at 09:33.

Hello Abdallahmaichine,

isn't the fact that since $ (u|u)_H\le(u,u)_V $ a Cauchy series in $ V $ is also one in $ H $ and the continuity of the scalar product in $ H $ showing the assertion ? (or am I overlooking something ?)

Best regards, Christian Schöner

Posted by CAS on 19 January 2015 at 23:51.

Hello Christian, yes every Cauchy sequence in V is a Cauchy sequence in H and then converge in H, but us we need convergence by the norm of V. The fact that V is a Hilbert space is equivalent to the closedness of the positive diagonal form a. You can also show that the associate linear diagonal operator is closed. (We lack something more!).

Bests regards, Abdallah Maichine.

Posted by Abdallahmaichine on 20 January 2015 at 11:46.

Dear Abdallah,

you are quite correct: it is not sufficient to have convergence in $ H $; but Christian's hint is the first step in the proof, namely how to get the candidate for the limit of the Cauchy sequence.

I will not carry it out here but I suggest you look at the proof of completeness of $ \ell_2 $. This should give you the idea how to do the remaining step. (Anyway, the case you have here is isometrically isomorphic to $ \ell_2 $ with weights; so you could use the completeness of such a space, if you are familiar with this.)

Best wishes, Jürgen

Posted by JürgenVoigt (administrator) on 21 January 2015 at 18:39.
Edited by JürgenVoigt (administrator) on 21 January 2015 at 18:40.

Dear professor Jürgen,

Ok, now I understand that the completeness of $\L^{2}(\mu)$, where $\mu$ is the weighted counting measure, and the isometry between H and $\l^{2}$ answers the quedtion.

Best regards, Abdallah.

Posted by Abdallahmaichine on 22 January 2015 at 13:41.

Misprint on line 3

[#21]

Dear all,

on line 3 of the lecture one should read $ (I-\Delta_{\rm D})^{-1} $. It is easy to get confused there. $ \Delta_{\rm D} $ is a negative self-adjoint operator, the `good' operator is the negative Dirichlet Laplacian.

Best wishes, Jürgen

Posted by JürgenVoigt (administrator) on 24 November 2014 at 13:26.
Edited by JürgenVoigt (administrator) on 24 November 2014 at 13:27.

Question on Example 6.4

[#20]

Dear ISem-Team

Firstly, I think there is a typo in the last line of the equivalences (a prime is missing). Secondly, I wonder whether the following argument for $ \bar{A} = -A^* $ works as well: We have $ A \subset -A^* $ as shown in the first part. The same argument (+ a density argument) should work to get $ A^* \subset -A^{**} $, hence $ A \subset -A^* \subset A^{**} = \bar{A} $. Clearly this implies $ \bar{A} = -A^* $.

Best wishes, Raffael

Posted by Raffael Hagger on 19 November 2014 at 17:11.
Edited by Raffael Hagger on 19 November 2014 at 17:12.

Dear Raffael,

you are speaking of the equivalences in Example 6.4. You are correct; the last equivalence should be $ -g'=f $; thanks! Concerning your question: `The same argument' would show that $ A^*|_{C_c^\infty}\subseteq-(A^*|_{C_c^\infty})^* $, and to conclude that this implies that $ A^*\subseteq A^{**} $ you would have to show that $ \overline{A^*|_{C_c^\infty}}=A^* $ (which just amounts to the statement one wants to show). In any case you end up with using the denseness of $ C_c^\infty(\mathbb R) $ in $ H^1(\mathbb R) $.

Best wishes, Jürgen

Posted by JürgenVoigt (administrator) on 19 November 2014 at 19:55.
Edited by JürgenVoigt (administrator) on 19 November 2014 at 20:11.

Dear Jürgen

Oh, I see. The thing I called density argument was just what you did below. The argument looked a bit too complicated to me when I first read it, but it was not. Thank you!

Best wishes, Raffael

Posted by Raffael Hagger on 20 November 2014 at 09:32.

Question c of exercice 6.3

[#49]

The eigenvalues of minus Laplacian in $ [0,\pi ] $ are $ (\lambda_n=n^{2}), n\in\mathbb{N} $ not $ (\lambda_n=n), n\in\mathbb{N} $. There is a mistake on the solution $u$ proposed in the exercise, the pulsation is $ \sqrt{\lambda} $ not $ \lambda $ and then you will one has $ sin(\sqrt{\lambda}\pi)=0 $ then $ \lambda_n=n^{2} $.

Abdallah Maichine.

Posted by Abdallahmaichine on 22 May 2015 at 19:39.