-
AN INTRODUCTION TO SOBOLEV SPACES
Steve Taylor, Montana State University.
Preface: These notes were written to supplement the graduate
level PDE course atMontana State University. Sobolev Spaces have
become an indispensable tool in the theoryof partial differential
equations and all graduate-level courses on PDE's ought to
devotesome time to the study of the more important properties of
these spaces. The object of thesenotes is to give a self-contained
and brief treatment of the important properties of Sobolevspaces.
The main aim is to give clear proofs of all of the main results
without writing anentire book on the subject! Why did I write these
notes? Much of the existing literature onthe subject seems to fall
into two categories, either long treatises on the subject with
themost general assumptions possible (and thus unsuitable for part
of a PDE course), or verysketchy discussions confined to a chapter
of a PDE text.
CONTENTS:
1. The Spaces W j , p () and W0j , p
().......................................... 12. Extension
Theorems...........................................................
103. Sobolev Inequalities and Imbedding
Theorems...................... 134. Compactness
Theorems...................................................... 205.
Interpolation
Results.........................................................
256. The Spaces H k() and
H0k()............................................. 277. Trace
Theorems................................................................
29Appendix: Some Spaces of Continuous
Functions.......................
34References................................................................................
35
In these notes, is a domain (i.e. an open, connected set) in Rn
.
1. The Spaces W j , p () and W0j , p ()
Definitions: Suppose 1 p < . Then(i) Llocp () = {u: u Lp (K)
for every compact subset K of }(ii) u is locally integrable in if u
Lloc1 () .(iii) Let u and v be locally integrable functions defined
in . We say that v is
the th weak derivative of u if for every C0()
1
-
uD dx = (1)| | v dx ,
and we say that Du = v in the weak sense.(iv) Let u and v be in
Llocp () . We say that v is the th strong derivative of u if
for each compact subset K of there exists a sequence { j} in C|
| (K) such that j u in Lp (K) and D j v in Lp (K).
THEOREM 1 If Du = v and Dv = w in the weak sense then D+ u = w
in the weaksense.
PROOF Let C0 () and = D . Then
uD + dx = (1)| | v dx = (1)| | vD dx = (1)| |+| | w dx .
Definition (mollifiers): Let C0 (Rn ) be such that(i) Supp B1(0)
, (recall that "supp" denotes the support of a function, and
Br (p) denotes an open ball of radius r and center p).(ii) (x)
dx =1 ,(iii) (x) 0 .If > 0 then we set (provided that the
integral exists)
Ju(x ) =1 n
( x y
)u(y) dy .
Ju is called a mollifier of u. Note that if u is locally
integrable in and if K is acompact subset of then Ju C
(K) provided that < dist(K,). Suppose now thatu Lloc
p () . Clearly
Ju(x ) = (y)u(x y) dyB1(0) ,so for p > 1 we have (if 1 / p +1
/ q = 1)
| Ju(x)| {(y)}1/ q{(y)}1/ p |u(x y)| dyB1 (0)
2
-
( ({(y)}1/ q )q dx)1/ q ( ({ (y)}1/ p| u(x y)|)p dy)1/ pB1(0)B1
(0) .
Hence | Ju(x)|p (y)|u(x y)|p dyB1 (0) , and this trivially holds
if p = 1 too. Integratingthis, we see that
| Ju(x)|pK dx (y) |u(x y)|p dx dyKB1(0) (y) |u(x)|p dx dy
K0B1 (0)= |u(x)|p dx
K0 ,where K0 is a compact subset of , K Interior(K0 ) and <
dist(K,K0 ) . i.e. we have
|| Ju||L p (K) ||u||Lp (K0 ) . (1)
LEMMA 2 If u Llocp () and K is a compact subset of then || Ju
u||Lp (K ) 0 as 0 .
PROOF Let K0 be a compact subset of where K Interior(K0 ) and
let < dist(K,K0 ) . Let > 0 and let w C(K0 ) be such that ||
u w||Lp (K0 ) < . Thenapplying (1) to u w , we obtain
|| Ju Jw||Lp (K ) < . (2)
But J w(x) w(x) = (y){w(x y) w(x)} dyB1(0 ) , and this goes to
zero uniformly on Kas 0. Hence, if is sufficiently small, we
have
|| Jw w||Lp (K ) < . (3)
Hence, by (2) and (3)
|| Ju u||Lp (K ) ||w u||Lp ( K) +|| Ju Jw||Lp (K ) +|| Jw w||Lp
( K) < 3 . (4)
Since is arbitrary, || Ju u||Lp (K ) 0 as 0.
3
-
The proof of the following theorem contains some other important
approximatingproperties of mollifiers.
THEOREM 3 Suppose that u and v are in Llocp () . Then Du = v in
the weak sense if and
only if Du = v in the strong Lp sense.
PROOF Suppose that Du = v in the strong Lp sense. Let C0() and
letK = supp . Let > 0 and take C| |(K) so that || u||Lp ( K)
< and|| D v||Lp (K ) < . Then
| uD dxK (1)| | v dxK || D dxK (1)| | D dxK |
+ | (u )D dx|K +| (v D ) dxK |
||u ||Lp ( K) || D ||Lq (K ) +||v D ||Lp ( K) || ||Lq (K ) (||D
||Lq (K ) +|| ||Lq (K ) ),
where q is the conjugate exponent of p (if p = 1 then q = and if
p > 1 then1 / p +1 / q = 1). But is arbitrary, so the LHS must
be zero. So Du = v in the weaksense.
Conversely, suppose that Du = v in the weak sense and let K be a
compact subsetof . Then Ju C
(K) if < dist(K,) and we have for all x in K
D Ju(x) = n Dx(x y
)u(y) dy
= n (1)| | Dy(
x y
)u(y) dy
= n (x y
)v(y) dy
= Jv(x).
But by Lemma 2, || Ju u||Lp (K ) 0 and || D Ju v||Lp ( K) =|| Jv
v||Lp ( K) 0 as 0.Thus Du = v in the strong sense.
Definitions (i) |u| j, p = ( | D u(x)| p dx| | j )1/ p .(ii) C
j, p () = {u C j(): | u| j, p < }.(iii) H j , p () = completion
of C j, p () with respect to the norm | | j, p .
4
-
H j , p () is called a Sobolev space. We will encounter other
such spaces as well.Recall that the completion of a normed linear
space is a larger space in which all Cauchysequences converge (i.e.
it is a Banach space). It is constructed by first defining a space
ofequivalence classes of Cauchy sequences. Two Cauchy sequences
{xm}, {ym} are said tobe in the same equivalence class if lim
m || xm ym ||= 0 . A member x of the old space is
identified with the equivalence class of the sequence {x, x,x, .
. .} of the new space and inthis sense the new space contains the
old space. Further, the old space is dense in itscompletion.
Moreover, if a normed linear space X is dense in a Banach space Y,
then Y isthe completion of X.
Recall that for 1 p < , Lp () is the completion of C0() with
respect to theusual "p norm". This knowledge allows us to see what
members of H j , p () "look like".Members of Lp () are equivalence
classes of measurable functions with finite p norms,two functions
being in the same equivalence class if they differ only on a set of
measurezero.
Suppose that {um} is a Cauchy sequence in C j, p (). Then for ||
j , {D um} is aCauchy sequence in Lp (). Hence, there are members u
of Lp () such that Dum uin Lp (). Hence, according to our
definition of strong derivatives, u0 is in Lp () and uis the strong
derivative of u0 . Hence we see that
H j , p () = {u Lp (): u has strong Lp derivatives of order j in
Lp () and there exists asequence {um} in C j, p () such that Dum Du
in Lp ()}.
Definition W j , p () = {u Lp (): the weak derivatives of order
j of u are inLp ()}
Note that by Theorem 3, an equivalent definition of W j , p ()
is obtained by writing "strongderivatives" instead of "weak
derivatives". Because of this, we see easily thatH j , p () W j ,
p() . In fact, H j , p () = W j, p (). This is not obvious because
formembers of H j , p () we can find sequences of nice functions
such that Dum Du inthe topology of Lp (), while according to our
definition of strong derivatives, such limitsexist only in the
topology of Lloc
p () for members of W j , p () . Before proving thatH j , p () =
W j, p (), we need the concept of a partition of unity.
5
-
LEMMA 4 Let E Rn and let G be a collection of open sets U such
that
E {U: U G} . Then there exists a family F of non-negative
functions f C0(Rn ) suchthat 0 f (x) 1 and
(i) for each
f F , there exists
U G such that supp f U ,(ii) if K E is compact then supp f K is
non-empty for only finitely many
f F ,(iii)
f (x)f F =1 for each x E (because of (ii), this sum is
finite),(iv) if
G ={1, 2 , . . .} where each i is bounded and i E then the
familyF of such functions can be constructed so that
F = {f 1, f 2, . .} andsupp f j j .
The family of functions F is called a partition of unity
subordinate to the cover G.
PROOF Suppose first that E is compact, so there exists a
positive integer N such thatE i =1
N Ui , where each
Ui G . Pick compact sets Ei Ui such that E i =1N Ei . Letgi = J
i Ei , where i is chosen to be so small that supp gi Ui . Then gi
C0
(Ui ) andgi > 0 on a neighborhood of Ei . Let g = gii =1
N , and let S = supp g i =1N Ui . If < dist(E,S) then k = J S
is zero on E and h = g + k C(Rn ). Further, h > 0 on Rnand h = g
on E. Thus
F = {f i : f i = gi / h} does the job.If E is open, let
Ei = E B i (0) {x: dist(x,E) 1i}.
Thus Ei is compact and E = i =1N Ei . Let
Gi be the collection of all open sets of the formU [Interior(Ei
+1) Ei 2 ], where
U G and E0 = E1 = . The members of
Gi providean open cover for the compact set Ei Interior(Ei1 ),
so they possess a partition of unity
F i with finitely many elements. We let
s(x) = g(x)gF ii =1
and observe that only finitely many terms are represented and
that s > 0 on E. Now we letF be the collection of all functions
of the form
6
-
f (x) =g(x)s(x) , x E
0, x E
This F does the job.If E is not open, note that any partition of
unity for U is a partition of unity for
E.For the proof of (iv), let H be the partition of unity
obtained above and let f i = sum
of functions h in H such that supp h i , but supp h j , j < i
. Note that each h isrepresented in one and only one of these sums
and that the sums are finite since each i isa compact subset of E.
Thus the functions f i provide the required partition of unity.
THEOREM 5 (Meyers and Serrin, 1964) H j , p () = W j, p ().
PROOF We already know that H j , p () W j , p() . The opposite
inclusion followsif we can show that for every u W j ,p and for
every > 0 we can find w C j, p such thatfor || j , || D w Du||L
p () < .
For m 1 let
m = {x : || x||< m, dist(x,) > 1m
}
and let 0 = 1 = . Let { m} be the partition of unity of part
(iv), Theorem 4,subordinate to the cover {m+ 2 m}. Each u m is j
times weakly differentiable and hassupport in m +2 m . As in the
"conversely" part of the proof of Theorem 3, we can pick
m > 0 so small that wm = J m (um ) has support in m +3 m 1
and |wm um | j, p < 2m .Let w = m =1
wm . This is a C
function because on each set m +2 m we havew = wm 2 + wm 1 + wm
+ wm +1 + wm +2 . Further,
|| D w Du||L p () =||m =1 D (wm u m)||L p () m =1
|| D (wm u m)||Lp () m =1
/ 2m = .
7
-
Remarks(i) The proof shows that in fact C() C j, p () is dense
in W j , p () .(ii) Clearly members of C() C j, p () are not
necessarily continuous on oreven bounded near . It would be very
useful to have the knowledge thatC( ) C j , p () or C j( ) C j, p
() is also dense in W j , p () . But the followingexample shows
that this cannot always be expected.
Problem 1 Let = {(x, y) : 1 < x2 + y2 < 2, y 0 if x >
0}, i.e. an annulus minus thepositive x-axis. Let w(x, y) = , the
angular polar coordinate of (x,y). Clearly w is inW1,1() because it
is a bounded continuously differentiable function. Show that we
cannotfind a C1( ) such that |u |1,1 < 2pi . (Note that is the
whole annulus).
The reason for the failure of the domain in Problem 1 is that
the domain is on eachside of part of its boundary. The following
definition expresses the idea of a domain lyingon only one side of
its boundary.
Definition A domain has the segment property if for each x there
exists anopen ball U centered at x and a vector y such that if z U
then z + ty for0 < t < 1.
We will not need the following theorem, so we don't prove it.
For a proof, see Adam'sbook. However, see Lemma 9 for the simpler
version of the result that we will need.
THEOREM 6 If has the segment property then the set of
restrictions to of functionsin C0
(Rn ) is dense in Wm, p () .
THEOREM 7 Change of Variables and the Chain Rule. Let V, be
domainsin Rn and let T: V be invertible. Suppose that T and T 1
have continuous, boundedderivatives of order j . Then if u W j ,p
() we have
v = u o T W j , p (V) and thederivatives of v are given by the
chain rule.
PROOF Let y denote coordinates in and let x denote coordinates
in V( y = T(x) ). If f Lp() then
foT Lp(V ) because
| foT|p dxV = | f |p J dy const. | f |p dy (5)
8
-
(Here J is the Jacobian of T 1 ).If u W j ,p (), let {um} be a
sequence in C j, p () converging to u in W j , p () and
set
vm = umoT . By the chain rule, if || j
Dxvm = (Dyum )oT R ,
where the R , are bounded terms involving T and its derivatives.
But for || jDy
u Lp ()
(Dyu)oT Lp (V) (Dyu)oTR , Lp (V) since the R , are
bounded.Further,
|| Dx vm (Dyu)oTR , ||Lp (V ) =|| (Dyum Dyu)oTR , ||Lp (V )
||(Dyum Dyu)oTR , ||Lp (V )
const. ||(Dyum Dyu)oT||Lp (V ) const. || Dyum Dyu||Lp ()
by (5). So ( = 0 case),
vm v = uoT in Lp (V ) and
Dxvm (Dyu)oT R , in
Lp (V ). This shows that v W j, p (V) and
Dxv = (Dyu)oT R , .
Definition W0j , p () =completion of C0() with respect to the
norm | | j, p .
Remarks (i) Clearly W0j , p () W j , p() because C0() C j ,
p().(ii) Saying that f W0j, p () is a generalized way of saying
that f and itsderivatives of order j 1 vanish on . e.g. W01, p ()
W2, p () is auseful space for studying solutions of the Dirichlet
problem for secondorder elliptic PDE's.(iii) C0j() W0j, p ()
because if f C0j(), we know that if issufficiently small then J f
C0 () and J f f in | | j, p norm.
Problem 2 Show that W j , p (Rn ) = W0j, p (Rn ) . Hint: Why is
it enough to show that C j, p (Rn ) W0j , p (Rn ) ?
Problem 3 Show that if is a domain in Rn , f W0j, p () and if f
is extended to bezero outside then the new function is in W j , p
(Rn ).Problem 4 Show that if y C1[0,1] and y(0) = y(1) = 0 then y
W01, p (0,1) . Use thisfact to show that for any f Lp(0,1) there is
a unique y W01, p (0,1) W2, p(0,1) such thaty"y = f . Hint: Solve
the problem first with f C0(0,1) and then take limits.
9
-
2. Extension Theorems
Most of the important Sobolev inequalities and imbedding
theorems that we willderive in the next section are most easily
derived for the space W0
j , p () which (seeProblem 3) can be viewed as being a subspace
of W j , p (Rn ). Direct derivations of theseresults for the spaces
W j , p () are tedious and difficult because of the boundary
behaviorof the functions (Adams uses the direct derivation approach
in his book). In this section weinvestigate the existence of
extension operators that allow us to extend functions inW j , p ()
to be functions in W j , p (Rn ). This will allow us to easily
deduce the Sobolevimbedding theorems for the spaces W j , p () from
the corresponding results for W j , p (Rn ).
LEMMA 8 Let u Rn and f Lp(Rn ) . Set f (x) = f (x +u) . Then lim
0 f = f inLp (Rn ) .
PROOF Given > 0, let C0(Rn ) be such that || f ||L p < .
Since uniformly on a sufficiently large ball containing the
supports of all (say, for 1), wecan pick so small that || ||Lp <
. Then
|| f f ||Lp || f ||L p +|| ||Lp +|| f ||L p < 3 .
LEMMA 9 Let R+n
= {x Rn : xn > 0} . C(R +n ) C j, p (R+n ) is dense in W j ,
p (R+n ).
PROOF Suppose f is in W j , p (R+n ) let > 0 and pick C(R+n )
C j, p (R+n ) so that|| D D f ||Lp ( R+n ) < for all || j . We
take the vector of Lemma 8 to beu = (0,0,0, . . ,1) and define
functions Lp (Rn ) as
(x ) = D (x) , xn > 00 , xn 0
Observe that for each > 0, C(R +n ) C j, p (R+n ) . By Lemma
8, we can pick > 0so that, for all || j , || ||Lp ( Rn ) < .
But this implies that || D D ||Lp (R+n ) < .Hence
|| D D f ||Lp ( R+n ) || D D ||Lp (R+n ) +||D D f ||Lp ( R+n )
< 2 .
10
-
LEMMA 10 There exists a linear mapping E0: Wj, p (R+n ) W j, p
(Rn) such that E0 f = f
in R+n and |E0 f | j , pRn C| f | j, pR+n , where C depends on
only n and p.
PROOF If f C(R +n ), define
E0 f (x) =f (x) , xn 0
ck f (x1, x2, . . , xn 1, kxn )k =1j +1 , xn < 0
where the constants ck are chosen so that E0 f (x) C j (Rn ) ,
i.e.
(k)m ckk =1j+1 =1, m = 0,1,2, . . , j .
It is easy to check that there is a constant C depending on only
n and p such that
|| D E0 f ||Lp (Rn ) C|| D f ||Lp ( R+n ) . (6)
If now f W j, p (R+n ) , take a sequence f m C(R +n ) C j, p
(R+n ) converging to f inW j , p (R+n ) (we can do this by Lemma
9). Then f m is a Cauchy sequence and (6) impliesthat E0 f m is a
Cauchy sequence in W j , p (Rn ). We denote the limit by E0 f .
Since|| D E0 f m ||Lp (Rn ) C|| D f m ||Lp (R+n ) , taking limits
shows that f satisfies (6).
Definition A domain is of class Cm if can be covered by bounded
open sets jsuch that there are mappings j : j B , where B is the
unit ball centered at the originand
(i) j( j ) = B R+n(ii) j( j ) = B R+n(iii) j Cm( j) and j1 Cm(B
).
(Because of (iii), all derivatives of order m of j and its
inverse are bounded).
THEOREM 11 If is a bounded domain of class Cm then there exists
a bounded linearextension operator E:Wm , p() Wm , p (Rn ) .
11
-
PROOF Since is compact (boundaries are always closed), we might
as wellassume that the number of sets j covering is a finite number
N . Let U = j =1N j andlet d = dist(,U ). Setting 0 = {x : dist(x,)
> d / 2} , we see that0,1,2 , . . , N cover . These sets also
cover , which is compact, so by the firstpart of the proof of Lemma
4, there exists a finite partition of unity 0,1,2 , . . ,N for and
supp j j . Recall that the support of a function is the closure of
the set on whichthat function is non-zero. Hence, supp j is even
bounded away from j .
Let f Wm, p () . Then f j Wm, p ( j ), so by our chain rule
theorem(Theorem 7)
w j = ( f j ) o j1 Wm, p (R+n B). Clearly supp w j is bounded
away fromB , so we can extend w j to be a member of Wm, p (R+n ) by
letting it be zero in R+n B . Wecan further extend w j to all of
R
n by use of the extension operator E0 of Lemma 10. Let
w j = E0w . If
-
3. Sobolev Inequalities and Imbedding Theorems
THEOREM 12 If Rn satisfies the cone condition (with height h and
opening ) and ifp > 1, mp > n then Wm, p () CB() and there is
a constant C depending on only , h,n and p such that for all u Wm ,
p() , sup|u| C|u|m , p .
Note: does not have to be bounded as Friedman suggests in his
Theorem 9.1!
PROOF Initially, suppose that u is in C m, p (). Let g C(R) be
such that g( t) =1if t 1 / 2 and g( t) = 0 if t 1. Let x and let
(r, ) denote polar coordinates centeredat x. Here, = (1, 2 , . . ,
n1 ) denotes the angular coordinates and we can describe thecone
with vertex x in polar coordinates as Vx ={(r,) : 0 r h, A}.
Clearly, we have
u(x) = r {g(r / h)0h u(r,)} dr
=
(1)m(m 1)! r
m 1 mrm {g(r / h)0
h u(r, )} dr ,after m-1 integrations by parts. Next, we
integrate with respect to the angular measure dS ,noting that the
left-hand-side becomes a constant times u(x).
u(x) = c r m 1 m
r m {g(r / h)0h u(r, )}drdSA
= c rm n m
rm {g(r / h)0h u(r, )} r n1drdSA
= c rm n m
r m {g(r / h)u(r, )}dVV x .Applying Hlder's inequality to this,
we obtain
|u(x)| const.||r m n ||Lq (V x ) || m
rm {g(r / h)u(r,)}||Lp (V x ) const. || rm n||Lq (V x ) |u|m, p
.
But r m n is in Lq(Vx) if n 1+ (m n)q > 1, which is the case
because q = pp1 andmp > n . Thus, we obtain sup|u| C|u|m , p .
To extend this result to arbitrary u Wm , p(),take a sequence {uk}
of functions in C m, p () converging to u in the | |m, p norm.
Thensup|u j uk | C|u j uk |m , p , showing that the sequence is a
Cauchy sequence in CB().
13
-
Thus u is in CB() and taking the limit of sup|u j | C|u j |m , p
shows that u satisfies the sameinequality.
Problem 5. Modify the proof to show that the theorem also
applies to the case p = 1,m = n .
Problem 6. Show that a similar theorem holds for W0m, p () and
the cone condition is
not required. Note that here we can even conclude that W0m, p ()
{u CB ( ) : u = 0 on
}.
COROLLARY 13 If Rn satisfies the cone condition (with height h
and opening ) and if p > 1, (m k)p > n then Wm, p () CBk()
and there is a constant C dependingon only , h, n, k and p such
that for all u Wm , p() sup| |k | D
u| C|u|m , p .
PROOF Apply the previous theorem to the derivatives Du for || k
.
Problem 7 What can you conclude if p = 1 and m k = n ? See
Problem 5.Problem 8 What is the corresponding theorem for W0
m, p () ? See Problem 6.
THEOREM 14 If Rn is any domain and p > n then W01, p () C0, (
) , where =1 np and there exists a constant C depending on only p
and n such that for allu W0
1, p()
|u(x) u(y)||| x y|| C || Diu||Lp ()i=1
n .
PROOF Let u C0 () . We might as well assume that u C0 (Rn ) .
Let
d =|| x y|| , Sx = Bd (x), Sy = Bd (y) and S = Sx Sy. Then
|u(x) u(y)| vol(S) = |u(x) u(y)| dzS
|u(x) u(z)|+|u(z) u(y)| dzS
|u(x) u(z)| dzSx + |u(z) u(y)| dzSy
But if (r, ) are the polar coordinates of z in a coordinate
system centered at x, we get|u(x) u(z)| 0r | u | d , which
implies
14
-
|u(x) u(z)| dzSx | u | d rn 1drdS0
r0d | u | d r
n1drdS0d0d
=
dn
n| u | d0
d dS=
dn
n1 n| u |
n 1d0
d dS=
dn
n1 n| u | dzSx
dn
n||1n ||Lq (Sx ) ||
u ||Lp (Sx )
where q = pp 1
. A simple calculation shows that
||1 n||Lq (Sx ) = const. d1
n
p
and it is easy to see that
|| u ||Lp (Sx ) const. || Diu||Lp ()i =1n .
Also, vol(S) = const. dn and the integral over Sy can be
estimated in a similar fashion.Putting this together yields
|u(x) u(y)| Cd1n
p || Diu||Lp ()i=1
n
which is precisely the inequality that we wanted. Further, we
know from Theorem 12applied to Rn that sup|u| C|u|1, p . Combining
this with the previous inequality shows thatfor u C0
() we have || u||C 0, ( ) C|u|1, p . Thus, if we now let u W01,
p() and take asequence {um} of functions in C0() converging to u in
| |1,p norm, it follows that {um}converges in C0, ( ) . Thus u C0,
( ) , and taking limits shows that u satisfies theinequality in the
statement of the theorem.
15
-
THEOREM 15 If Rn is any domain and p < n then W01, p () Lr ()
, wherer =
npn p
and there exists a constant C depending on only p and n such
that for allu W0
1, p()
|| u||Lr () C || Diu||L p ()i =1
n .
Remark The proof relies on a simple generalization of Hlder's
inequality which canbe proved by induction by using Hlder's
inequality. The inequality states
|u1u2u3 . . um | dx ||u1||L p1 ||u2||Lp2 . . ||um ||Lpm (7)where
1
p1+
1p2
+1p3
+ . . . +1pm
= 1.
PROOF of Theorem 15. It suffices to prove the result for u
C01(Rn ). First we prove
the result for the case p = 1. For each i we have
|u(x)| | Diu| dxi
xi | Diu| dxi
.Multiplying these n inequalities together and taking the n 1 th
root gives
|u(x)|n
n 1 ( | Diu| dxi
) 1n1i =1
n (8)
Observe that | Diu| dxi
does not depend on xi , but it does depend on all n 1 of
theremaining variables. We integrate each side of (8) with respect
to x1 and use thegeneralized Hlder inequality with pi = m = n 1 to
obtain
|u(x)|n
n1 dx1
( | D1u| dx1
) 1n 1 ( | Diu| dxi
) 1n 1i= 2
n dx1
( | D1u| dx1
) 1n1 ( | Diu| dxi
dx1
) 1n1i = 2
n .
16
-
The RHS is still a product of n 1 functions of x2 , so we
integrate each side with respectto x2 , again applying (7) with pi
= m = n 1. Continuing in this manner, we finally obtain
|u(x)|n
n1 dxRn ( | Diu| dxRn
i =1
n )1
n1
i.e. || u||L
n
n1 ( | Diu| dxRn
i =1
n )1n (geometric mean)
1n
|Diu| dxRni=1
n (arithmetic mean)
Here we have used the fact that an arithmetic mean is no less
than a geometric mean of thesame numbers. This proves the result
for the case p = 1.
For p > 1, let = (n 1)pn p
= 1 + n(p 1)n p
. Since > 1 and u C01(Rn ), it follows
that |u| C01(Rn) . Clearly
Di |u| = (n 1) pn p
|u|n( p1)
n p (Diu) .
We apply the p = 1 case to |u| and obtain
( |u|np
n p dxRn )
n 1n
1n
(n 1)pn p
|u|n( p1)
n p | Diu| dxRni =1
n
(n 1)pn(n p) ( (|u|
n( p1)n p )
pp1 dx
Rn )p1p || Diu||Lp
i =1
n=
(n 1)pn(n p) ( |u|
npn p dx
Rn )p1
p || Diu||L pi =1
n
Hence
( |u|np
n p dxRn )
n pnp
(n 1) pn(n p) || Diu||Lpi =1
n ,
which is the desired result. As usual, to obtain the same result
for a function u W01, p(),
we just take a sequence of functions in C01 converging to u.
17
-
Remark According to the theorem, W01, p () Lr (), where r is
given above. But
obviously W01, p () Lp (), so by the following interpolation
lemma, W01, p () Lq ()
for all q satisfying p q r . If is bounded then clearly this
holds for all q satisfying1 q r .
LEMMA 16 If s q r and Ls () Lr () , then Lq () and
|| ||Lq || ||Ls || ||Lr1 ,
where = s(r q)q(r s) .
PROOF Apply Hlder's inequality to the integral of | |q , using
the facts that| |(1 )q L
r
(1 )q and | | q L
s
q.
Problem 9 Modify the proof of Theorem 15 to show that if p = n
>1 then
W01, p () Lr () for every r p . Hint: First prove this for the
case r > n
2
n 1 by setting
= r n 1n
and applying the p = 1 result to |u| . The r nn 1
norm that shows up on the
RHS after applying Hlder's inequality (as we did in our proof
above) can be estimated interms of the n = p norm and the r norm by
use of Lemma 16. Finally, obtain the result forall r p by applying
Lemma 16 to the result that you have just proved.
COROLLARY 17 For every domain in Rn there exists a constant C
depending ononly n and p such that
a) if kp < n then W0k, p () Lnp
n kp () and for each u W0k , p ()
|| u||L
npnkp
C|u|k , p
b) if kp > n then W0k, p () Cm , ( ) , where m is the integer
satisfying0 < k m n
p
-
PROOF a) If || k 1 and u W0k , p then Du W01, p , which is
contained in Lnp
n p
by Theorem 15. Thus W0k, p
W0k 1,
npn p
. Iterating this process once more, we find
W0k, p
W0k 2,
npn2 p
. Continuing the iterations culminates in the desired
result.
b) Since (k m 1)p < n , case (a) implies that W0k m 1, p
Lnp
n (k m1)p= L
n
(1 ). Thus if
u W0k , p
and || m +1, then Du W0k m 1, p . Hence W0k, p W0m +1, n(1 )
. But this shows
that if u W0k , p
and || m then Du W01, n(1 )
, which is contained in C0, ( ) byTheorem 14. Thus W0
k, p () Cm , ( ) .
Remarks A few "particular cases" have been left out because they
require separateproofs (see Problem 10 below). They are:
i) If kp = n and p > 1 then W0k, p () Lq () for all q
satisfying p q < .ii) If kp = n and p = 1 (so that k = n) then
W0k, p () CB ( ) .
iii) If kp > n , p > 1 and np
is an integer then W0k, p () W0
k n
p,q
() for all q satisfying
p q < .
iv) If kp > n and p = 1 (so np
is obviously an integer) then W0k, p () CBk n ( ).
All of the particular cases listed above have the appropriate
norm inequalities associatedwith them.
Problem 10 Use the results of Problems 5, 6, 7, 8 and 9 to prove
the particular caseslisted above.
COROLLARY 18 If is a bounded C1 domain in Rn (or any other
domain such thatthere exists a bounded extension operator E : W1, p
() W1, p (Rn) ) then the statementsconcerning the spaces W0
k, p () in Corollary 17 and in the remark following the
corollaryalso apply to the spaces W k, p () . However, the constant
C may also depend on .
19
-
PROOF The cases for k = 1 dealt with in Theorems 14 and 15 are
easily seen tohave their counterparts here because of the extension
operator. Inspection of the proof ofCorollary 17 shows how the
results for k >1 may be derived from the results for k =
1without any additional assumptions on the domain.
Remark One can show that extension operators exist for Lipschitz
domains and evendomains satisfying certain cone conditions (see the
remarks following the proof ofTheorem 11). This Sobolev imbedding
theorem is thus valid for such domains.
Definition Let A and B be Banach spaces. If A B , we say that A
is continuouslyimbedded in B (in symbols, this is written
A\o(\s\do3( ), )B)if there is a constant Csuch that || x||B C||
x||A for all x A .
The theorems in this section provide examples of imbeddings and
are called
Sobolev Imbedding Theorems. e.g. W01, p () \o(\s\do3( ), ) L
npn p
for p < n . It is easy to see that A\o(\s\do3( ), )B is
equivalent to the identity mapping fromA into B being continuous
(i.e. bounded).
4. Compactness Theorems
Definition Suppose that A\o(\s\do3( ), )B. We say that A is
compactly imbeddedin B if every sequence bounded in A has a
subsequence that converges in B.
e.g. If K is compact then any bounded sequence in C1(K) is a set
of equicontinuousfunctions so, by the Arzela-Ascoli Theorem, it has
a subsequence that converges in C(K) .i.e. C1(K) is compactly
imbedded in C(K) .
Recall that if A and B are Banach spaces and if M : A B is a
bounded linearmapping then M is said to be compact if for every
bounded sequence {xm} in A thesequence {Mxm} has a subsequence that
converges. Thus, saying that A is compactlyimbedded in B is
equivalent to saying that the identity mapping from A into B is
compact. Itis easy to see that if M : A B and P : B C are bounded
linear mappings and A, B andC are Banach spaces then PM is compact
if one of the mappings A or B is compact.Consequently, we obtain
the very useful result that if A\o(\s\do3( ), )B and B\o(\s\do3(),
)C then the imbedding A\o(\s\do3( ), )C is compact if one of the
other twoimbeddings is compact.
20
-
LEMMA 19 Suppose that is a bounded domain. Ifa) 0 < 1 then
Cm, ( ) is compactly imbedded in Cm( ).b) 0 < < 1 then Cm, (
) is compactly imbedded in Cm, ( ) .
PROOF It suffices to prove the results for m = 0 because, once
this is done, we canapply this case to the derivatives of the
functions and deduce the result for general m.. Let{ f j} be a
sequence in C0, ( ) such that || f j ||C0, M . But this implies| f
j(x) f j (y)| M || x y|| , showing that the sequence is a bounded,
equicontinuous setof functions. By the Arzela-Ascoli Theorem, there
exists a subsequence { f jk } thatconverges in C( ). Thus C0, ( )
is compactly imbedded in C( ).
We show below that the same subsequence also converges in C0, (
). Supposethat C0, ( ). Then
[ ]0, = sup| (x) (y)|
||x y|| = sup| (x) (y)|
|| x y||
| (x) (y)|1
21
([ ]0, )
(max | | )1
We apply this to f jk f j r , noting that [ f j k f jr ]0, [ f
jk ]0, + [ f jr ]0, 2M , and obtain
[ f j k f jr ]0, 2M
(max | f jk f jr | )1
,
showing that the subsequence is a Cauchy sequence in C0, ( )
(because it converges inC( )). Thus the subsequence converges in
C0, ( ).
COROLLARY 20 If is bounded, kp > n and 0 < k m np
-
COROLLARY 21 If is a bounded C1 domain (or any other domain for
which thereis a bounded extension operator E : W1, p () W1, p (Rn)
), kp > n and 0 < k m n
p
-
(ii) Every sequence in E has a convergent subsequence.(iii) E is
totally bounded.
The Theorem gives us two other very useful characterizations of
compact mappings andimbeddings.
THEOREM 23 If is bounded and p < n, then W01, p () is
compactly imbedded inLq() for all q < np
n p.
PROOF Consider first the case q = 1. Let A be a bounded set in
W01, p () . We may
consider the members of A as members of W1, p (Rn ) with
supports contained in . LetAh = {Jhu : u A}. Note that we have
| Jhu(x)| h n ( x zh )|u(z)| dz h n (max)|| u||L1 ()and | Di
Jhu(x)| hn 1 | Di( x zh )||u(z)| dz h n1(max| Di|)||u||L1 () .Since
is bounded, || u||L1 const.||u||L p . The inequalities above show
that Ah is abounded equicontinuous set of functions in C( ). By the
Arzela-Ascoli Theorem, everysequence in Ah has a subsequence that
converges in C( ). Obviously, such subsequencesalso converge in
L1(), so we see that Ah is totally bounded in L1().
If u A then
u(x) Jhu(x) = (z)(u(x) u(x hz)) dz|z|1
= (z) r u(x rz
|| z|| ) dr0h| |z|| dz
|z|1 .
Thus |u(x) Jhu(x)| (z) | Diu(x r z|| z|| )|i =1n dr0
h ||z || dz|z |1 .
Integrating this with respect to x, we find
23
-
|u(x) Jhu(x)| dx (z) | Diu(x r z||z||)|Rni =1n dx dr0
h ||z || dz|z |1
= (z) | Diu(x)|i =1
n dx dr0h| |z|| dz
|z|1
= (z)h||z|| | Diu(x)|i=1
n dx dz|z|1
h |Diu(x)|i=1
n dx hB, (9)
where B is a constant depending on our bound of members of A in
W01, p () (again we use
the fact that the L1 norm is weaker than the Lp norm on a
bounded domain).Let > 0. Since Ah is totally bounded in L
1(), we can cover Ah by a finitenumber of balls Bi of radius /
2. Let h =
2B. By (9), if Jhu Bi , then u is contained in
a ball of radius centered at the center of Bi . Thus, A is
covered by a finite number ofballs of radius . i.e. A is totally
bounded in L1(). Thus W01, p () is compactly imbeddedin L1().
Suppose W01, p (). Then Lnp
n p by Theorem 15 and we get from Lemma 16
(with s = 1 and r = npn p
) that
|| ||Lq || ||L1 || ||L
npn p
1 C|| ||L1 (|| || Di ||L pi =1
n )1
Now let {um} be a bounded sequence in W01, p () and assume |um
|1, p M . Since W01, p ()is compactly imbedded in L1(), we can
extract a subsequence {um j} that converges inL1(). Applying the
inequality above to um j umk , noting that |um j umk |1, p 2M ,
weobtain
|| um j umk ||Lq const.||um j umk ||L1 ,
showing that the subsequence is a Cauchy sequence in Lq() .
Hence the subsequenceconverges in Lq() and W01, p () is compactly
imbedded in Lq() .
24
-
COROLLARY 24 If kp < n and is bounded then W0k, p () is
compactly imbeddedin Lq() for all q < np
n kp.
PROOF W0k, p () is continuously imbedded in W0
1,np
n (k 1) p (), which is compactlyimbedded in Lq() if q <
np
n kp, by Theorem 23.
COROLLARY 25 The same compactness results hold for W k, p () if
is a bounded,C1 domain (or any other type of bounded domain for
which there is an extension operatorE : W1, p () W1, p (Rn) .
PROOF See the proof of Corollary 21.
Remark The case kp = n is missing from the previous results. But
since W k, p iscontinuously imbedded in W k, r for all r < p
(provided that the domain is bounded), itfollows from Corollary 25
that W k, p is compactly imbedded in Lq() for all q < . Thesame
applies to W0
k, p ().
5. Interpolation Results
The following results are very useful in PDE theory. We make use
of Theorem 26in our proof of Grding's Inequality in our study of
elliptic problems.
THEOREM 26 Let u W0k , p () . Then for any > 0 and any 0
-
First suppose that u C02 (R) and consider an interval (a,b) of
length b a = . If
y (a,a + / 3) and z (b / 3,b), then by the Mean Value Theorem
there is ap (a,b) such that
| u ( p)|=| u(z ) u(y)z y
| 3
(|u(z)|+|u(y)|)
Consequently, for every x (a,b), we obtain
| u (x)|=| u ( p) + u (t) dtp
x | 3 (|u(z)|+|u(y)|) + | u (t)| dtab .
Integrating with respect to y and z over the intervals (a, a + /
3) and (b / 3,b)respectively, we obtain
| u (x)| | u (t)| dta
b + 18 2 |u(t)| dtab ,
so by Hlder's inequality and the inequality (A + B)p 2p1(A p + B
p ),
| u (x)|p 2p 1({ | u (t)| dta
b }p + (18)p 2p { | u(t)| dtab }p )
2p1({ | u (t )|p dta
b }{ 1 dtab }p1 + (18)p
2 p{ |u(t)|p dt
a
b }{ 1 dtab }p1 )= 2p1 ( p1 | u (t)|p dt
a
b + (18)p p+1 |u(t)|p dtab ).
Integrating this with respect to x over the interval (a,b)
gives
| u (x)| p dxa
b = 2 p1( p | u (t)|p dtab + (18)p
p|u(t)|p dt
a
b ).We now subdivide R into intervals of length and obtain by
adding all of theseinequalities that
| u (x)|p dx
2p 1( p | u (t)| p dt
+ (18)p p |u(t)|p dt ) (11)
26
-
Suppose now that u C0 (Rn ) . Then we can apply (11) to u
regarded as a function of xi
and integrate with respect to the remaining variables to
obtain
| uxi |p dx
Rn 2p1( p | 2u
x i2|p dx
Rn + (18)p
p|u|p dx
Rn )Taking the pth root of this and using (Ap + Bp )1/ p A + B,
we obtain (10). (Actually, wedon't quite obtain (10). We actually
obtain the inequality (10) for 2 instead of . Butsince is an
arbitrary positive constant, (10) holds). Finally, (as usual) to
obtain the resultfor u W0
(), we take a sequence of functions in C0 converging to u.
COROLLARY 27 The interpolation inequality stated in Theorem 26
also applies tomembers of W k, p () , provided that is a bounded C2
domain (or any other domain forwhich there is a bounded extension
operator E : W2, p () W2,p (Rn ) . Here the constant Cmay also
depend on p and .
PROOF Because of the extension operator, an inequality of the
form (10) holds forfunctions in W2, p (). The full result follows
by induction from this case.
6. The Spaces H k() and H0k().
Definitions H0k() = W0k, 2() and H k() = W k, 2(). These spaces
are Hilbert spaces
with inner product
(u,v)k = Du(x)Dv (x) dx|| k .
If = Rn , we get a very useful representation of such functions
in terms of theFourier-Plancherel Transform. Recall that for
functions f L2 (Rn ) , we define the Fourier-Plancherel transform
of f as
f () = limR
(2pi ) n/ 2 eix . f (x) dx||x | | R .
27
-
The limit exists in the topology of L2(Rn ), || f ||L2 =|| f
||L2 , and f can be recovered by usingthe inversion formula
f (x) = limR
(2pi ) n/ 2 e ix . f ( ) d| | || R .
Again, the limit here exists in the topology of L2(Rn ). The
reason for the limits in theseformulas is that the integrands are
not necessarily in L1(Rn ). Clearly, the Fourier-Plancherel
transform is an isometric (i.e. norms are equal) isomorphism (i.e.
boundedlinear mapping with a bounded inverse) from L2(Rn ) onto
L2(Rn ).
Integration by parts shows that for f C0(Rn ), the transform of
fx j
is i j f ( ).From this, we see by induction that the transform
of D f is (i ) f (). If now we letf H k(Rn ) and take a sequence of
C0 functions converging to f, we find that (i ) f ()is in L2(Rn )
for all || k and the transform of D f is (i ) f (). Thus, we see
that iff H k(Rn ) then f H k(Rn ), where
H k(Rn ) ={g L2 (Rn ) : (1+|| ||) k g( ) L2(Rn )}.
It is easy to see that C0
is dense in H k and if g is in C0
then g is the transform of aninfinitely differentiable, rapidly
decaying function f (a function in the Schwarz class, to
beprecise). Taking limits, we see that if g is in H k then g is the
transform of a function fbelonging to H k . Further, if we define
an inner product on H k as
(u,v) H k = 2u( )v () d| |k ,
we find that || f ||k =|| f || H k . Thus, the
Fourier-Plancherel transform is an isometricisomorphism from H k
onto H k . Questions about functions in H k are thus
transformedinto equivalent (and often simpler) questions about
functions in H k .
Problem 12 Consider the initial value problem for the wave
equation
d2udt 2
= u , u(0) = f H k , dudt
(0) = g H k 1 .
28
-
(We think of u as being a function of t taking values in H k ).
Construct a candidate u for asolution using Fourier transforms.a)
Show that u is a continuous H k -valued function of t.b) Show that
u is a continuously differentiable H k 1-valued function of t.c)
Show that du
dt is a continuously differentiable H k 2 -valued function of t
and that
d2udt 2
= u in H k 2 .
d) How large does k have to be in order for u to be a classical
(i.e. C2 ) solution.Hints: Clearly it suffices to answer the
equivalent questions about u . Use the DominatedConvergence Theorem
to help you answer a), b), c). For d), use the Sobolev
ImbeddingTheorem.
7. Trace Theorems.
In PDE Theory, one often needs to know how functions behave on
boundaries ofdomains. If f is a function defined on a domain , we
call the restriction of f to thetrace of f. If all we know about f
is that it is in some Lp space, then the trace of f is
notwell-defined because has measure zero. However, if kp > n and
is a bounded C1
domain in Rn , then we know by Corollary 18 that functions in W
k, p () are continuous on and thus they have well-defined traces
that are bounded functions. In this section, weconcern ourselves
with the important case kp < n.
In the following results, a vector x in Rn is denoted by x = ( x
, xn ) , where x belongs to Rn1.
LEMMA 27 If u W1,1 (Rn ) , then for every R , the function v( x
) = u( x , ) is inL1(Rn1) , and
|| v||L1( Rn1 ) ||u||L1 (Rn ) +|| Dnu||L1 (Rn ) .
Remark. One needs to be careful when talking about traces of
equivalence classes offunctions. The trace certainly exists for u
C0
(Rn ) . For u W1,1 (Rn ) , we know that wecan find a sequence of
functions in C0
(Rn ) that converges to u. The norm inequalityasserted in the
lemma shows that the sequence of traces of these functions
converges inL1(Rn1) . It is in this sense that the trace of u
exists in L1(Rn1) .
29
-
PROOF It suffices to prove the result for the case = 0 and u C0
(Rn ) . By theMean Value Theorem for integrals
|u( x , xn )| d x dxnRn1
0
1
= |u( x ,)| d x Rn1
for some [0,1]. But
|u( x ,0)|=|u( x , ) Dnu( x ,t) dt0
||u( x ,)|+ | Dnu( x ,t)| dt
0
1
.
Integrating this over Rn1 gives
|| v||L1( Rn1 ) |u( x ,)| d x Rn1 + | Dnu( x ,t)| dt
0
1
Rn1 d x
= |u( x , )| d x Rn1
0
1
dt + | Dnu( x ,t)| dt0
1
Rn1 d x .
This completes the proof of the lemma.
LEMMA 28 If u W1, p(Rn ) where p < n, then for every R , the
functionv( x ) = u( x , ) is in Lr (Rn 1) , where
r =(n 1)pn p
= 1+ n(p 1)n p
and there is a constant C depending on only n and p such
that
|| v||Lr (Rn1) C |u|1, pRn
.
PROOF We can assume that p > 1 because the p = 1 case is
dealt with in theprevious lemma. We first show that if u W1, p(Rn )
then w =|u|r W1,1 (Rn ) and
|| w||L1 const. || Du||Lpr 1 ||u||Lp , || Diw||L1 const. ||
Du||Lpr . (12)
30
-
It suffices to prove this result for the case u C0 (Rn ) . Let q
= p / (p 1) . Then
(r 1)q = np / (n p) , so by the Sobolev Imbedding Theorem (Th.
15),
|| |u|r 1 ||Lqq const. ||Du||L pnp / (n p)
and combining this with Hlder's Inequality, we get the first of
(12):
|| w||L1 = |u|r dx = |u|r1 |u| dx ||u||Lp || |u|r 1 || Lq
const.|| Du||Lpr 1 ||u||Lp .Since Diw = r|u|r1 Diu , we obtain the
second of (12):
|| Diw||L1 = r|| |u|r 1 ||Lq || Diu||Lp const. || Du||L pr .
We now apply Lemma 27 to w and immediately obtain the
inequality
|| v||Lr (Rn1) const.(|| Du||Lpr1 ||u||L p +|| Du||Lpr )1/ r
const.(|| Du||Lp11/ r ||u||Lp1/ r +||Du||L p )
const.(||u||Lp +||Du||Lp ).
LEMMA 29 If u Wk , p (Rn ) where kp < n , then for every R ,
the functionv( x ) = u( x , ) is in Lr (Rn 1) , where
r =(n 1)pn kp
and there is a constant C depending on only n, k and p such
that
|| v||Lr (Rn1) C |u|k , pRn
.
PROOF By Sobolev's Imbedding Theorem (Th. 15) applied to the
first orderderivatives of u, we have u W1,np / (n(k 1)p )(Rn) . Now
apply Lemma 28.
31
-
Reminder: Parametrized Surface Integrals.
If X(u) = (x1(u1,u2 ), x2 (u1,u2 ), x3 (u1,u2 )) is a
parametrization for a smooth surfaceS in R3 , it is well-known from
elementary calculus that one may integrate functions definedon S
using the formula
f (x) dSS = f o X(u) K(u) du1du2
,
where is the domain of X and
K(u) =|| Xu1 Xu2
|| = (((x3, x2 )(u1,u2 ) )2 + ((x1,x3 )(u1,u2 )
)2 + ((x1, x2 )(u1,u2 ))2 )1/ 2 .
Differential Geometry yields a generalization of this formula.
Suppose now that is a domain in Rn1 and that X: Rn is a
parametrization for a smooth hypersurface S .Then surface integrals
over S may be calculated using
f (x) dSS = f o X(u) K(u) du
,
where
K(u) = ( ( (x1,x2,K, x k,K, xn )(u1,u2 ,K,un ))2
k =1
n )1/ 2 .
Here the x k notation means that the xk term does not
appear.
THEOREM 30 Suppose that is bounded and is of class C k . If u Wk
, p () wherekp < n , then the restriction v of u to is in Lr ()
, where
r =(n 1)pn kp
and there is a constant C depending on only n, k and p and such
that
|| v||Lr () C |u|k , p .
32
-
Remark. See the remark following the statement of Lemma 27 for
clarification of thephrase "restriction v of u to ". The same
remark applies because we know that anextension operator E: W k ,
p() Wk , p (Rn ) exists, and thus the restriction to offunctions in
C0
(Rn ) is dense in W k, p ().
PROOF Let E: W k , p() Wk , p (Rn ) be the extension operator of
Theorem 11.Since any u Wk , p () is associated with an element Eu
Wk , p (Rn ), we might as well juststudy the properties of the
trace on of C0(Rn ) functions.
Let j and j be as in the definition of Ck domains. Since is
compact, we
might as well assume that there is a finite number of the j , 1
j N , covering . Let j 1 j N , be a partition of unity for
subordinate to this cover. If u C0 (Rn ) , then
( ju)o j1 C0k (B) and we can extend
( ju)o j1 to be in C0k (Rn ) by defining thefunction to be zero
outside B. By Lemma 29, the trace w j of
( ju)o j1 on the hyperplaneP: yn = 0 satisfies
|| w j ||Lr (P ) C|( ju)o j1|k , pB Cj |u|k , pRn
,
where C depends on only n, p, and k and C j is independent of
u..
X j (y) = j1(y1,K,yn1,0) is a parametrization for the
hypersurface S j = () j andwe may estimate the trace
v j = wj o j of ju on this hypersurface using
thisparametrization (see the "reminder" preceding the statement of
the theorem).
|v j(x)|r dSS j = |w j (y)|r K j(y) dy
PB Rj | wj |r dy
P B ,
where Rj = max(K j ). Comparing this to the preceding
inequality, we see that there is aconstant M j independent of u
such that
|| v j ||Lr (S j ) M j |u|k, pRn
.
The function v satisfies a similar inequality because v = v j .
Finally, it is clear that theresult holds for arbitrary u Wk , p ()
(see the remark following the statement of thetheorem).
33
-
Problem 13 Modify the proof of Lemma 27 to show that if u W1,
p(Rn ), then for every R , the function v( x ) = u( x , ) is in Lp
(Rn1) , and there exists a constant Kdepending only on n and p such
that|| v||Lp (Rn1 ) K|u|1, pR
n
.
Problem 14 Deduce from the previous problem and Lemma 28 that
the function v ofLemma 28 belongs to Lq(Rn1 ) for all q satisfying
p q r .
Appendix: Some Spaces of Continuous Functions.
Here, we define the spaces of continuous functions that appear
in these notes. Caution:Notation and definitions of such function
spaces vary from text to text. Recall that westated that is a
domain in Rn . The connectedness of is not needed in the
followingdefinitions, so we need only assume that is an open subset
of Rn .
1. C() is the set of functions continuous in .2. C( ) is the set
of functions continuous in .3. C k () is the set of functions which
have derivatives of order k that are
continuous in .4. C k ( ) is the set of functions in C( ) which
have derivatives in of order k
that can be extended to be members of C( ).5. C() is the set of
functions in C k () for all k.6. C( ) is the set of functions in C
k ( ) for all k.7. C0() is the set of functions in C() that have
supports that are compact subsets
of (recall that the support of a function is the closure of the
set on which thefunction fails to vanish). Since is open, such
functions necessarily vanish in aneighborhood of the boundary of
.
8. C0k () is the set of functions in C k () that have supports
that are compact subsets
of .9. C0
() is the set of functions in C0k () for all k.10. CB() is the
set of bounded functions in C(). This is a Banach space when
equipped with the "sup norm".11. CB( ) is the set of bounded
functions in C( ). This is a Banach space when
equipped with the "sup norm". If is bounded, this space
coincides with C( ).
34
-
12. CBk () is the set of functions in CB() with derivatives of
order k belonging to
CB(). This is a Banach space if we define the norm of a member f
of this space assup|| k, x | D
f (x)|.13. CB
k ( ) is the set of functions in both CBk () and C k ( ). This
is a Banach space,equipped with the same norm as in (12). If is
bounded, this space coincides withC k ( ).
14. C k , ( ) , where 0 < 1, is the set of functions in CBk (
) that have derivatives oforder k that are uniformly Hlder
continuous with exponent . C k , ( ) is aBanach space with norm
|| f ||C k, = sup|| k, x | D f (x)|+[ f ]k, ,
where [ f ]k , = supx , y, x y, | |= k
| D f (x) D f (y)||| x y|| .
References
The results stated in these notes appear in most texts on
Sobolev spaces, includingthose listed below. However, there are
many different proofs of the results. For thisreason, the key
lemmas and theorems that appear in these notes are listed below
with areference to the source which has a proof that most resembles
the proof in these notes.
Th. 1 [Fr], Th. 6.1L. 2 [Fr], Th. 6.2Th. 3 [Fr], Th. 6.2L. 4
[Zi], L. 2.3.1Th. 5 [Fr], Th. 6.3; [Ad], Th. 3.16.Th. 6 [Ad], Th.
3.18.Th. 7 [Ad], Th. 3.35. (see [Zi], Th. 2.2.2 for Lipschitz
changes of variables).Th. 11 [Ad], Th. 4.26; [G.T.], Th. 7.25.Th.
12 [Fr], Th. 9.1Th. 14 [Fr], Th. 9.2Th. 15 [G.T.], Th. 7.10L. 19
[Ad], Th. 1.31
35
-
Th. 22 [Ru], Appendix A4Th. 23 [G.T.], Th. 7.22Th. 26 [G.T.],
Th. 7.27; [Ad], Th. 4.13
There is obviously no room here for a complete bibliography. For
a more complete list ofreferences, the reader should refer to the
bibliographies of the texts listed below.
Ad. R. A. Adams, Sobolev Spaces, Academic Press, 1975.
Fr. A. Friedman, Partial Differential Equations, Krieger,
1983.
G.T. D. Gilbarg and N. S. Trudinger, Elliptic Partial
Differential Equations of SecondOrder, Springer-Verlag, 1983.
Ma. V. G. Maz'ja, Sobolev Spaces, Springer-Verlag, 1985.
Ru. W. Rudin, Functional Analysis, MacGraw-Hill, 1973.
Zi. W. P. Ziemer, Weakly Differentiable Functions,
Springer-Verlag, 1989.
36