I am reading Functional Analysis, Sobolev Spaces and Partial Differential Equations, by Haim Brezis, and also Evans' PDE. I am confused about what $||\nabla u||_{W^{1,p}(\Omega)}$ precisely means. In these two books it is often written, like for example in the Sobolev, Gagliardo and Nirenberg inequality, which states that:
$$ || u||_{p^*} \leq C ||\nabla u||_p $$
The book defines $\nabla u=(u_{x_1},..., u_{x_n})$, so it is a vector. But the book also defines the L^p (And Sobolev $W^{1,p}$) spaces as those formed by functions from $\mathbb{R}^n$ to $\mathbb{R}$, and the norm by integrating that function.
Is there a standard meaning for $||\nabla u||_{W^{1,p}(\Omega)}$?
Because, for example, I have checked in the proof of $ || u||_{p^*} \leq C ||\nabla u||_p $ that it means that $||u||_p$ is bounded by the product of the $L^p$ norms of the derivatives of $u$. But if we check the proof of Proposition 9.3, we have, at a certain step, for $h \in \mathbb{R}^n$:
$$\int_0^1 |\nabla u(x+t)|^p dt$$
How does one interpret that integral? I understand taking the absolute value of a vector could mean take the absolute value of every component in that vector, but integrating a vector? Am I missing something?
Means that $\partial_{x_j} u \in L^p(\Omega)$ $\forall j=1,...,n$, i.e. $\nabla u \in L^p(\Omega) \times \cdot \cdot \cdot \times L^p(\Omega)$.