11

From PDE Evans, page 272. My question is towards the bootom of this post.

THEOREM 1 (Trace Theorem). Assume $U$ is bounded and $\partial U$ is $C^1$. Then there exists a bounded linear operator $$T : W^{1,p}(U) \rightarrow L^p(\partial U)$$ such that

$\quad$(i) $Tu=u|_{\partial U}$ if $u \in W^{1,p}(U) \cap C(\bar{U})$

and

$\quad$(ii) $$\|Tu\|_{L^p(\partial U)} \le C \| U \|_{W^{1,p} (U)},$$ for each $u \in W^{1,p}(U)$, with the constant $C$ depending only on $p$ and $U$.

DEFINITION. We call $Tu$ the trace of $u$ on $\partial U$.

Proof. 1. Assume first $u \in C^1(\bar{U})$. As in the first part of the proof of Theorem 1 in §5.4 let us also intially suppose $x^0 \in \partial U$ and $\partial U$ is flat near $x^0$, lying in the plane $\{x_n=0\}$. Choose an open ball $B$ as in the previous proof and let $\hat{B}$ denote the concentric ball with radius $r/2$.

$\quad$Select $\zeta \in C_c^\infty(B)$, with $\zeta \ge 0$ in $B$, $\zeta \equiv 1$ on $\hat{B}$. Denote by $\Gamma$ that portion of $\partial U$ within $\hat{B}$. Set $x'=(x_1,\ldots,x_{n-1}) \in \mathbb{R}^{n-1} = \{x_n=0\}$. Then \begin{align} \int_\Gamma |u|^p \, dx' &\le \int_{\{x_n=0\}} \zeta |u|^p \, dx' \\ &= -\int_{B^+} (\zeta |u|^p)_{x_n} \, dx \\ &= -\int_{B^+} |u|^p \zeta_{x_n} + p|u|^{p-1} (\text{sgn} u)u_{x_n} \zeta \, dx \\ &\le C \int_{B^+} |u|^p + |Du|^p \, dx, \end{align} where we employed Young's inequality, from §B.2.

The proof continues on in the textbook but I cut it short here. My question was how was Young's inequality used exactly?

For reference, the textbook's appendix (§B.2) says:

Young's inequality. Let $1 < p, q < \infty, \frac 1p + \frac 1q = 1$. Then $$ab \le \frac{a^p}{p}+\frac{b^q}{q} \quad (a,b > 0).$$

I have a strong suspicion that Young's inequality must be used in the last step, because it is one of the only two steps with the $\le$ sign (the other step with $\le$ is the very first one on top). I do know in the third line, the product rule was used to expand $(\zeta |u|^p)_{x_n}$.

K.defaoite
  • 13,890
Cookie
  • 13,952
  • I don't understand the first line in the proof. I really don't understand how Evans is able to say $$ \int_{{x_n=0}}\zeta |u|^p dx'=-\int_{B^+}D_n(\zeta |u|^p)dx$$ I am guessing this is some kind of integration by parts computation, but the details are really not clear. Could you help me? – K.defaoite Apr 28 '25 at 01:47

1 Answers1

12

Let $q = p/(p-1)$. Then $p,q$ satisfy the requirements of Youngs inequality, so that

$$ |u|^{p-1} |u_{x_n}| \leq C [|u|^{q(p-1)} + |u_{x_n}|^p]= C [|u|^p + |u_{x_n}|^p]. $$

The other terms (the signum and terms involving $\zeta$ can be bounded by constants).

PhoemueX
  • 36,211
  • Use Fubini on the right hand side to get something like $\int_{???} \int_0^\infty (\zeta |u|^p){x_n} d x_n d(x_1, \dots, x{n-1})$ and apply the fudamental theorem of calculus in the inner integral (noting that the $\zeta |u|^p$ vanishes at $\infty$, because $\zeta$ has compact support). – PhoemueX Jul 19 '14 at 08:40
  • Yes, I saw your hint but I was still thinking about it. The $???$ was throwing me off a little; still thinking what I need to fill in. So far I'm trying, according to your recommendation with Fubini's theorem: $$\int_{{x_n=0}} \zeta |u|^p , dx'=\int_{{x_n=0}} \int_0^\infty (\zeta |u|^p){x_n} , dx_n dx' = \int_0^\infty \int{{x_n=0}} (\zeta |u|^p){x_n} , dx' dx_n = \int{B^+} (\zeta |u|^p)_{x_n} dx_n $$ but I am so not sure about the last inequality (or any one of them, might as well). – Cookie Jul 20 '14 at 17:46
  • @Cookie I am reading the same proof and I am stuck in this step (Not the one the main question asks, but the one that seems to be discussed here) Did you conclude the steps are correct? $\zeta$ has compact support inside $B$, so... Did you use the formula $f(x)=f(a) + \int_a^x g(t)dt$? I still don't get why there is a minus before the integral. – D1X May 06 '16 at 17:16
  • How does one obtain the bound for every derivative $u_{x_i}$? Evans only shows it for $x_n$. Do you have to apply the same trick for every variable $x_i$? – D1X Jun 01 '16 at 16:12
  • @D1X: Evans assumes (in the part of the proof shown in the question) that the boundary lies in the hyperplane ${x_n =0} $. Thus, only the $x_n $ derivative is needed. If you assume instead that it lies in ${x_i =0} $, you have to apply the same trick w.r.t. the different variable. But I don't think this is needed (I don't have the book with me right now), Evans needs to reduce the general case to the special case considered here somewhere. – PhoemueX Jun 01 '16 at 16:38
  • Yes , but he concludes (Or seems to) $... \leq \int |u|^p +|Du|^p dx$ directly. I assume $Du$ means any derivative, but I can only conclude is $... \leq \int |u|^p +|u_{x_n}|^p dx$. – D1X Jun 01 '16 at 16:45
  • @D1X: $Du $ denotes the full gradient/Jacobian. Hence, it is clear that $|u_{x_n}|\leq |Du|$. – PhoemueX Jun 01 '16 at 17:01
  • Is the constant $C$ towards the end, i.e., $||u||^p_{L^{p}(\partial U)}\leq C||u||^p_{W^{1,p}(U)}$ dependent on the way we choose the open covering?. I presume the constant depends on $N$ - the number of finite open sub covers covering the boundary in which case the constant chosen is a 'bad' constant. – Alexander Jan 20 '17 at 09:53
  • @Alexander: The constant will depend on the number $N$ of balls in the covering, but also on the partition of unity $(\zeta_n)n$ subordinate to that covering: In the part of the proof which is stated in the original question, the derivative $\zeta{x_n}$ appears. Also, one uses a certain diffeomorphism to "flatten out" the boundary. The properties of these diffeomorphisms will also enter into the final constant. I do not know exactly what you mean by a "bad" constant. One which one does not know very well or over which one has essentially no control? – PhoemueX Jan 20 '17 at 14:57
  • @PhoemueX I have noticed the other factors on which the constant depend on. By a 'bad' constant I mean one which is not solely dependent on $U$, dimension $N$, $p$. As you rightly said that it has dependency on the test function and its derivative, open covers etc. which actually is a bad thing in my view. – Alexander Jan 20 '17 at 16:27
  • @Alexander: But all these things (covering, partition of unity and diffeomorphisms) only depend on $U $ :) – PhoemueX Jan 20 '17 at 19:08
  • @PhoemueX I don't understand the first line in the proof. I really don't understand how Evans is able to say $$ \int_{{x_n=0}}\zeta |u|^p dx'=-\int_{B^+}D_n(\zeta |u|^p)dx$$ I am guessing this is some kind of integration by parts computation, but the details are really not clear. Could you help me? – K.defaoite Apr 28 '25 at 01:47