I have three question regarding the appearance of the space of locally summable functions in the definition of weak derivatives and sobolev spaces.

The deifinition of weak derivatives from Evans:

Suppose $v, u \in L^1_{loc}(U)$, and $\alpha$ is a multiindex. We say that v is the $\alpha$:th weak partial derivative of $u$, written $D^{\alpha}u=v$, provided

$\int_U uD^\alpha\phi dx =(-1)^{|\alpha|}\int_U v\phi dx $, for all test functions $\phi \in C^\infty_c(U)$

1) Why is it necessary to use the space $L^1_{loc}(U)$ and not for example $L^1(U)$?

In the definition of a Sobolev space $W^{k,p}(U)$ in Evans:

The Sobolev space $W^{k,p}(U)$ consists of all locally summable functions $u$: $U \rightarrow \mathbb{R} $ such that for each multiindex $\alpha$ with $|\alpha| \leq k, D^\alpha u $ exists in the weak sense and belongs to $L^p(U)$.

As I understand, locally summable means that $u \in L^1_{loc}(U)$.

2) Does the usage of $L^1_{loc}(U)$ in this definition comes from the fact that it is used in the definition of the weak derivative above?

Last, the Sobolev norm $\|u\|_{W^{k,p}(U)}$ is defined as:

$\|u\|_{W^{k,p}(U)} = (\sum_{ |\alpha| \leq k}\int_U|D^\alpha u|^p dx)^{\frac{1}{p}}$,

3) But for a function $u \in W^{k,p}(U)$ how do we now that the term $\int_U|u|^p dx$ (that appears in the Sobolev norm when $\alpha = (0,0, ...,0)$) is $<\infty$ since all we know is that $u \in L^1_{loc}(U)$ and we dont have that $u \in L^p (U)$?

Thanks :)


I had a similar question while reading through Evans. I don't know the answer to all your questions, but I think I can respond to some of them.

"As I understand, locally summable means that $u \in L^1_{loc}(U)$."

Yes, this is correct. One can verify by looking at the Wikipedia page for locally summable functions, for example.

As for your question 2, I think this is true. It's a bit strange to me because most books I have seen do not have the locally summable part, but maybe it's a bit more general.

As for your question 3, the full definition of a Sobolev space in the book is that the Sobolev space $W^{k, p}(U)$ consists of all locally summable functions $u: U \to \mathbb{R}$ s.t. for each multiindex $\alpha$ with $|\alpha| \leq k$, $D^\alpha u$ exists in the weak sense and belongs to $L^p(U)$.

So taking $\alpha = 0$ in the definition is exactly what you want.


The answer to 1) is that in order to make sense of the expression $\int u \phi$ for $\phi$ a $C^\infty$ function with compact support, it is enough that $u \in L_\mathrm{loc}^1$ (you do not need that $u \in L^1$).

What is the added value of allowing $u$ to be only in $L_\mathrm{loc}^1$ instead of asking that $u \in L^1$? Well, you want to be able to say that weak derivatives are a generalization of usual derivatives. But $C^1$ functions are not in general in $L^1(\mathbb R^n)$, though they are always in $L_\mathrm{loc}^1(\mathbb R^n)$. So if you want to talk about the weak derivative $u(x)=x^2$ on $\mathbb R$, and prove that it is equal to $v(x)=2x$, you need to work with $L_\mathrm{loc}^1(\mathbb R)$.


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.