2
$\begingroup$

As part of a proof of the Picard–Lindelöf theorem, I am using the following space:

$X = \{ u \in C([0,T]) : u(0) = \alpha , || u - \alpha || \leq K\}$

where $K \in \mathbb{R}_{> 0} , \ \alpha \in \mathbb{R}$ and the norm is defined as follows:

$|| u || = ^{\text{sup}}_{t \in (0,T)} |u(t)|$.

I have read and understood the proof just fine, but no version I have read so far has provided a proof for whether or not $X$ is actually complete, which is necessary to make use of the Banach Fixed Point Theorem.

I have shown in previous exercises that such $X$ is complete under the norm:

$||u|| = ^{\text{sup}}_{t \in (0,T)} ( |u(t)| + |u'(t)| ) $

but I cannot figure out how to do so without the derivative included in the norm.

Here is my progress so far:

Let $u_{n}$ be a Cauchy sequence in $X$. I have shown that there is a function, $u$, to which $u_{n}$ converges pointwise, and thus $u$ is continuous. I have also shown that $||u|| \leq K$. However, I have not figured out how to show that $u$ is differentiable, and thus inside $X$.

I would like to prove the following claim:

if $u_{n}$ is a Cauchy sequence in $X$, then $u_{n}'$ is also Cauchy in $X$.

I would surely be done then. Can someone guide me as to how I can prove this claim? Or if by any chance I am completely mistaken here and $X$ is in fact not complete at all? In that case, how does one make use of such a set to prove the Picard–Lindelöf theorem?

Thank you very much.

$\endgroup$
  • $\begingroup$ Your $X$ is a subset of the normed space $C([0,T])$, but it is linear space itself. $\endgroup$ – daw Mar 19 at 12:33
1
$\begingroup$

Your $X$ does not consist of differentiable functions, but continuous ones (obviously all differentiable functions are in there). There is no need to even talk about $u'$. More to the point your space $X$, with that norm, is only "interested" in continuous functions.

You state that $u_n$ converges to $u$ pointwise, thus $u$ is continuous, but this is not true. You need to show that $u_n$ converges to $u$ uniformly, and only then can you conclude that $u$ is continuous. However this follows quite simply from the fact that $u_n$ must converge to $u$ in the $\sup$ norm because $u_n$ is Cauchy in the $\sup$ norm and converges pointwise to $u$.

$\endgroup$
  • $\begingroup$ Hi K. Power, Thank you very much! I realise now that I simply misunderstood the notation of C([0,T])... This makes my job much easier indeed. Thank you. $\endgroup$ – David Hughes Mar 19 at 8:47
  • $\begingroup$ Hi K.Power, if you wouldn't mind me asking, I'm wondering where I can find help for the last part of this question. I have shown that a solution, u, to the Integral Equation is unique in X, and I would like to show that it is unique in all C([0,T]). $\endgroup$ – David Hughes Mar 20 at 17:19
  • $\begingroup$ @DavidHughes Are you asking where to find a proof of the Picard-Lindelöf Theorem? We have shown $X$ is complete so you can use it in the proof $\endgroup$ – K.Power Mar 20 at 17:23
  • $\begingroup$ I am asking about one final point of the Picard Lindelof theorem. I understand how u is unique in X, which is a subset of C([0,T]) where the functions are bounded. I do not know how to show that this uniqueness carries over to all C([0,T]) $\endgroup$ – David Hughes Mar 20 at 17:24
  • $\begingroup$ @DavidHughes I'd recommend just asking a new question here on MSE, and linking to this post. Make sure to give context to the question and make clear exactly what you aren't clear about. $\endgroup$ – K.Power Mar 20 at 17:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.