I wanted to show, that $p>1$ in Aubin-Lions Lemma is necessary. I hoped there already is an example for $X=Y=Z=\mathbb{R}$ so that the embedding \begin{align} \{u \in L^\infty(0,T);\mathbb{R}) \mid u^\prime \in L^1(0,T;\mathbb{R})\} \to C([0,T];\mathbb{R}) \end{align} is not compact.

So, I have to find a sequence of bounded functions in $\{u \in L^\infty(0,T);\mathbb{R}) \mid u^\prime \in L^1(0,T;\mathbb{R})\}$ that don't have a converging subsequence in $C([0,T];\mathbb{R})$.

My problem is that I didn't get an idea of which type of bounded functions I'm actually searching for. At first, I thought I have to take a sequence of oscillating functions like $\sin(kx)$ but their derivative isn't bounded. Does anybody have an idea or a hint?


$\DeclareMathOperator{AC}{AC}$Notice that the requirement the space you are considering is precisely the space of absolutely continuous functions. Thus, the question reduces to

Can we find a bounded sequence of absolutely continuous which does not converge uniformly to a continuous function?

In particular, notice that it would be sufficient to take a sequence $u_n$ of absolutely continuous functions which converges pointwise to a discontinuous function. There are many examples, for instance $$u_n(x) = \frac{x^n}{T^{n+1}}$$

It seems worth pointing out that in the setting you are considering, the Aubin-Lions lemma is the same as the humble Rellich-Kondrakov theorem. In particular, the space $$A = \{u \in L^\infty((0,T); \mathbb{R}) | u' \in L^1(0,T;\mathbb{R})\}$$ is equivalence (as a Banach space) to $W^{1,1}((0,T))$, since on one hand, $L^1((0,T)) \subseteq L^\infty((0,T))$ by Holder's inequality, while on the other hand $L^\infty((0,T)) \subseteq W^{1,1}((0,T))$ by Poincare's inequality.


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.