Fernandez
Fernandez
Fernandez
1. Introduction
When dealing with nonlinear elliptic equations with critical growth (in the sense
of the Sobolev embeddings) the concentration - compactness principle by Lions, see
[12], have been proved to be a fundamental tool for proving existence of solutions.
Just to cite a few references, we have [1, 2, 3, 7, 4, 11] but there is an impressive
list of references on this topic.
Recently in the analysis of some new models, that are called electrorheological
fluids, the following equation has been studied
−∆p(x) u = f (x, u) in Ω. (1.1)
The operator ∆p(x) u := div(|∇u|p(x)−2 ∇u) is called the p(x)-Laplacian. When
p(x) ≡ p is the well-known p-Laplacian.
In recent years a vast amount of literature that deal with the existence problem
for (1.1) with different boundary conditions (Dirichlet, Neumann, nonlinear, etc)
have appeared. See, for instance [5, 6, 8, 13, 14] and references therein.
However, up to our knowledge, no results are available for (1.1) when the source
term f is allowed to have critical growth at infinity (see the remark after the
introduction for more on this). That is,
|f (x, t)| ≤ C(1 + |t|q(x) )
with q(x) ≤ p∗ (x) := N p(x)/(N − p(x)) (if p(x) < N ) and {q(x) = p∗ (x)} = 6 ∅.
This article attempts to begin filling this gap. So, the objective is to extend the
concentration - compactness principle by Lions to the variable exponent setting.
The method of the proof follows the lines of the ones in the original work of
P.L. Lions and the main novelty in our result is the fact that we do not require the
exponent q(x) to be critical everywhere. Moreover, we show that the delta masses
are concentrated in the set where q(x) is critical.
Theorem 1.3. Let p(x) and q(x) be as in Theorem 1.1 and let r(x) be continuous.
Moreover, assume that maxΩ p < minΩ r and that there exists η > 0 such that
r(x) ≤ p∗ (x) − η in Ω.
Then, there exists λ0 > 0 depending only on p, q, r, N and Ω, such that if
inf λ(x) > λ0 for some δ > 0,
x∈Aδ
1,p(x)
problem (1.6) has at least one nontrivial solution in W0 (Ω). Here, Aδ is the
δ-tubular neighborhood of A, namely
Aδ := ∪x∈A (Bδ (x) ∩ Ω).
Organization of this article. After finishing this introduction, in Section 2 we
give a very short overview of some properties of variable exponent Sobolev spaces
that will be used throughout the paper. In Section 3 we deal with the main result
of the paper. Namely the proof of the concentration - compactness principle (The-
orem 1.1). In Section 4, we begin analyzing problem (1.6) and prove Theorem 1.3.
Finally, in Section 5, we prove Theorem 1.2.
Comment on a related result. After this paper was written, we found out
that a similar result was obtained independently by Yongqiang Fu [10]. Even the
techniques in Fu’s work are similar to the ones in this paper (and both are related
to the original work by Lions), we want to remark that our results are slightly
more general than those in [10]. For instance, we do not require q(x) to be critical
everywhere (as is required in [10]) and we obtain that the delta functions are located
in the criticality set A (see Theorem 1.1).
Also, in our application, again as we do not required the source term to be
critical everywhere, so the result in [10] is not applicable directly. Moreover, in
Theorem 1.3 our approach allows us to consider λ(x) not necessarily a constant
and the restriction that λ is large is only needed in an L∞ -norm in the criticality
set.
We believe that these improvements are significant and made our result more
flexible that those in [10].
1,p(x)
Define W0 (Ω) as the closure of C0∞ (Ω) with respect to the W 1,p(x) (Ω) norm.
1,p(x)
The spaces Lp(x) (Ω), W 1,p(x) (Ω) and W0 (Ω) are separable and reflexive Banach
spaces when 1 < inf Ω p ≤ supΩ p < ∞.
As usual, we denote p0 (x) = p(x)/(p(x) − 1) the conjugate exponent of p(x).
Define (
N p(x)
∗ if p(x) < N
p (x) = N −p(x)
∞ if p(x) ≥ N .
The following results are proved in [9].
0
Proposition 2.1 (Hölder-type inequality). Let f ∈ Lp(x) (Ω) and g ∈ Lp (x) (Ω).
Then the following inequality holds
Z
|f (x)g(x)| dx ≤ Cp kf kLp(x) (Ω) kgkLp0 (x) (Ω) .
Ω
The following proposition is also proved in [9] and it will be very useful here.
Proposition 2.5. Set ρ(u) := Ω |u(x)|p(x) dx. For u, ∈ Lp(x) (Ω) and {uk }k∈N ⊂
R
and so
lim sup kφkLq(x) (Ω) ≤ kφkLq(x) (Ω) .
νj ν
j→∞
Let λ0 := lim inf j→∞ kφkLq(x) (Ω) and assume that λ0 < kφkLq(x) (Ω) .
νj ν
Now we need a lemma that is the key role in the proof of Theorem 1.1.
Lemma 3.2. Let µ, ν be two non-negative and bounded measures on Ω, such that
for 1 ≤ p(x) < r(x) < ∞ there exists some constant C > 0 such that
kφkLr(x) (Ω) ≤ CkφkLp(x) (Ω)
ν µ
Then, there exist {xj }j∈J ⊂ Ω and {νj }j∈J ⊂ (0, ∞), such that
ν = Σνi δxi
For the proof of the lemma above, we need a couple of preliminary results.
Lemma 3.3. Let ν be a non-negative bounded measure. Assume that there exists
δ > 0 such that for all A Borelian, ν(A) = 0 or ν(A) ≥ δ. Then, there exist {xi }
and νi > 0 such that X
ν= νi δxi
The proof of the above lemma is elementary and is omitted.
Lemma 3.4. Let ν be non-negative and bounded measures, such that
kψkLr(x) (Ω) ≤ CkψkLp(x) (Ω)
ν ν
Then there exist δ > 0 such that for all A Borelian, ν(A) = 0 or ν(A) ≥ δ.
Proof. First, observe that if ν(A) ≥ 1,
Z Z
χA (x) p(x) χA (x) p(x)
1 dν ≤ 1 dν = 1.
Ω ν(A) p− Ω ν(A) p(x)
1
Then ν(A) p− ≥ kχA kLp(x) . On the other hand,
ν
Z Z
χA (x) r(x) χA (x)
1 dν ≥ dν = 1.
Ω ν(A) r+ Ω ν(A)
1
Then ν(A) r+ ≤ kχA kLr(x) . So we conclude that
ν
1 1
ν(A) r+ ≤ Cν(A) p− .
EJDE-2010/141 CONCENTRATION-COMPACTNESS PRINCIPLE 7
Proof of Lemma 3.2. By reverse Hölder inequality (3.2), the measure ν is absolutely
continuous with respect to µ. As consequence there exists f ∈ L1µ (Ω), f ≥ 0, such
that ν = µbf . Also by (3.2), we have
1 1 1 1
min ν(A) r− , ν(A) r+ ≤ C max µ(A) p− , µ(A) p+
for any Borel set A ⊂ Ω. In particular, f ∈ L∞ µ (Ω). On the other hand the
Lebesgue decomposition of µ with respect to ν gives us
µ = νbg + σ, where g ∈ L1ν (Ω), g ≥ 0
and σ is a bounded positive measure, singular with respect to ν.
Now consider (3.2) applying the test function
1
φ = g r(x)−p(x) χ{g≤n} ψ.
We obtain
1
kg r(x)−p(x) χ{g≤n} ψkLr(x)
ν
1
≤ Ckg r(x)−p(x) χ{g≤n} ψkLp(x)
µ
1
= Ckg r(x)−p(x) χ{g≤n} ψkLp(x)
gdν+dσ
r(x) 1
≤ Ckg p(x)(r(x)−p(x)) χ{g≤n} ψkLp(x) + Ckg r(x)−p(x) χ{g≤n} ψkLp(x)
ν σ
Since σ ⊥ ν, we have
1 r(x)
kg r(x)−p(x) χ{g≤n} ψkLr(x) ≤ Ckg p(x)(r(x)−p(x)) χ{g≤n} ψkLp(x)
ν ν
r(x)
Hence calling dνn = g (r(x)−p(x)) χg≤n dν the following reverse Hölder inequality holds
kψkLr(x) ≤ CkψkLp(x) .
νn νn
8 J. FERNÁNDEZ BONDER, A. SILVA EJDE-2010/141
By Lemma 3.3 and Lemma 3.4, there exists xni and Kin > 0 such that νn =
r(x)
n n r(x)−p(x) ν. Then, the points xn are in fact
P
i∈I Ki δxi . On the other hand, νn % g i
independent of n, and there will denoted by xi , and the numbers Kin are monotone
in n. Then, we have
r(x) X
g r(x)−p(x) ν = Ki δxi
i∈I
r(xi )
where Ki = g r(xi )−p(xi )
(xi )ν(xi ). This finishes the proof.
The following Lemma follows exactly as in the constant exponent case and the
proof is omitted.
Lemma 3.5. Let fn → f a.e and fn * f in Lp(x) (Ω) then
Z Z Z
lim |fn |p(x) dx − |f − fn |p(x) dx = |f |p(x) dx
n→∞ Ω Ω Ω
Let us now show that the points xj actually belong to the critical set A. In fact,
assume by contradiction that x1 ∈ Ω \ A. Let B = B(x1 , r) ⊂⊂ Ω − A. Then
q(x) < p∗ (x) − δ for some δ > 0 in B and, by Proposition 2.2, The embedding
W 1,p(x) (B) ,→ Lq(x) (B) is compact. Therefore, uj → u strongly in Lq(x) (B) and
so |uj |q(x) → |u|q(x) strongly in L1 (B). This is a contradiction to our assumption
that x1 ∈ B.
Now we proceed with the proof. Applying (3.1) to φuj and taking into account
that uj → u in Lp(x) (Ω), we have
SkφkLq(x) (Ω) ≤ kφkLp(x) (Ω) + k(∇φ)ukLp(x) (Ω) .
ν µ
Consider φ ∈ Cc∞ (Rn ) such that 0 ≤ φ ≤ 1, φ(0) = 1 and supported in the unit
ball of Rn . Fixed j ∈ I, we consider ε > 0 be arbitrary.
We denote by φε,j (x) := ε−n φ((x − xj )/ε). By decomposition of ν, we have:
Z
ρν (φi0 ,ε ) := |φi0 ,ε |q(x) dν
Ω
Z X
= |φi0 ,ε |q(x) |u|q(x) dx + νi φi0 ,ε (xi )q(xi ) ≥ νi0 .
Ω i∈I
For the rest of this article, we will denote
+ −
qi,ε := sup q(x), qi,ε := inf q(x),
Bε (xi ) Bε (xi )
p+
i,ε := sup p(x), p−
i,ε := inf p(x).
Bε (xi ) Bε (xi )
EJDE-2010/141 CONCENTRATION-COMPACTNESS PRINCIPLE 9
Then 1 1
+ −
q q
i,ε i,ε
min{νi , νi }S ≤ kφi,ε kLp(x) (Ω) + k(∇φi,ε )ukLp(x) (Ω) .
µ
By Proposition 2.5,
− +
k(∇φi,ε )ukLp(x) (Ω) ≤ max{ρ((∇φi,ε )u)1/p ; ρ((∇φi,ε )u)1/p }.
Then, by Hölder inequality, we have
Z
ρ((∇φi,ε )u) = |∇φi,ε |p(x) |u|p(x) dx
Ω
≤ k|u|p(x) kLα(x) (Bε (xi )) k|∇φi,ε |p(x) kLα0 (x) (Bε (xi )) ,
where α(x) = n/(n − p(x)) and α0 (x) = n/p(x).
Moreover, using that ∇φi,ε = ∇φ x−x
1
ε , we obtain
i
ε
+ −
k|∇φi,ε |p(x) kLα0 (x) (Bε (xi )) ≤ max{ρ(|∇φi,ε |p(x) )p /n
; ρ(|∇φi,ε |p(x) )p /n
},
and
Z
p(x)
ρ(|∇φi,ε | )= |∇φi,ε |n dx
Bε (xi )
x − xi n 1
Z
= |∇φ( )| n dx
Bε (xi ) ε ε
Z
= |∇φ(y)|n dy.
B1 (0)
Lp(x) (U ) for all U ⊂ Ω. By weakly lower semi continuity of norm we obtain that
dµ ≥ |∇u|p(x) dx and, as |∇u|p(x) is orthogonal to µ1 , we conclude the desired
result. This completes the proof.
4. Applications
In this section, we apply Theorem 1.1 to study the existence of nontrivial solu-
tions of the problem
−∆p(x) u = |u|q(x)−2 u + λ(x)|u|r(x)−2 u in Ω,
(4.1)
u = 0 on ∂Ω,
where r(x) < S p∗ (x) − ε, q(x) ≤ p∗ (x) and A = {x ∈ Ω : q(x) = p∗ (x)} = 6 ∅. We
define Aδ := x∈A (Bδ (x) ∩ Ω) = {x ∈ Ω : dist(x, A) < δ}. The ideas for this
application follow those in [11].
For (weak) solutions of (4.1) we understand critical points of the functional
|∇u|p(x) |u|q(x) |u|r(x)
Z
F(u) = − − λ(x) dx
Ω p(x) q(x) r(x)
4.1. Proof of Theorem 1.3. We begin by proving the Palais-Smale condition for
the functional F, below certain level of energy.
1,p(x)
Lemma 4.1. Assume that r ≤ q. Let {uj }j∈N ⊂ W0 (Ω) a Palais-Smale se-
1,p(x)
quence then {uj }j∈N is bounded in W0 (Ω).
Proof. By definition F(uj ) → c and F 0 (uj ) → 0. Now, we have
1 1
c + 1 ≥ F(uj ) = F(uj ) − hF 0 (uj ), uj i + hF 0 (uj ), uj i,
r− r−
where Z
hF 0 (uj ), uj i = |∇uj |p(x) − |uj |q(x) − λ(x)|uj |r(x) dx.
Ω
Then, if r(x) ≤ q(x), we conclude that
Z
1 1 1
c+1≥ − |∇uj |p(x) dx − |hF 0 (uj ), uj i|.
p+ r− Ω r−
We can assume that kuj kW 1,p(x) (Ω) ≥ 1. As kF 0 (uj )k is bounded we have that
0
1 1 − C
c+1≥ − kuj kp 1,p(x) − kuj kW 1,p(x) (Ω) .
p+ r− W 0 (Ω) r− 0
1 1
Let us show that if c < p+ − −
qA
S n and {uj }j∈N is a Palais-Smale sequence,
with energy level c, then I = ∅. In fact, suppose that I 6= ∅. Then let φ ∈ C0∞ (Rn )
with support in the unit ball of Rn . Consider, as in the previous section, the
rescaled functions φi,ε (x) = φ( x−xε ).
i
1,p(x)
As F 0 (uj ) → 0 in (W0 (Ω))0 , we obtain that
lim hF 0 (uj ), φi,ε uj i = 0.
j→∞
1 1
≥ − − Sn.
p+ qAδ
As δ is positive and arbitrary, and q is continuous, we have
1 1
c≥ − − Sn.
p+ qA
Therefore, if
1 1
c< − − Sn,
p+ qA
the index set I is empty.
Now we are ready to prove the Palais-Smale condition below level c.
1,p(x)
Theorem 4.2. Let {uj }j∈N ⊂ W0 (Ω) be a Palais-Smale sequence, with energy
1 1
n 1,p(x)
level c. If c < p+ − q− S , then there exist u ∈ W0 (Ω) and {ujk }k∈N ⊂
A
1,p(x)
{uj }j∈N a subsequence such that ujk → u strongly in W0 (Ω).
Proof. We have that {uj }j∈N is bounded. Then, for a subsequence that we still
denote {uj }j∈N , uj → u strongly in Lq(x) (Ω). We define F 0 (uj ) := φj . By the
1,p(x)
Palais-Smale condition, with energy level c, we have φj → 0 in (W0 (Ω))0 .
0 1,p(x)
By definition hF (uj ), zi = hφj , zi for all z ∈ W0 (Ω); i.e.,
Z Z Z
|∇uj |p(x)−2 ∇uj ∇z dx − |uj |q(x)−2 uj z dx − λ(x)|uj |r(x)−2 uj z dx = hφj , zi.
Ω Ω Ω
Then, uj is a weak solution of the following equation.
−∆p(x) uj = |uj |q(x)−2 uj + λ(x)|uj |r(x)−2 uj + φj =: fj in Ω,
(4.5)
uj = 0 on ∂Ω.
1,p(x) 1,p(x)
We define T : (W0 (Ω))0 → W0 (Ω), T (f ) := u where u is the weak solution
of the equation
−∆p(x) u = f in Ω,
(4.6)
u = 0 on ∂Ω.
Then T is a continuous invertible operator.
1,p(x)
It is sufficient to show that fj converges in (W0 (Ω))0 . We only need to prove
1,p(x)
that |uj |q(x)−2 uj → |u|q(x)−2 u strongly in (W0 (Ω))0 . In fact,
Z
h|uj |q(x)−2 uj − |u|q(x)−2 u, ψi = (|uj |q(x)−2 uj − |u|q(x)−2 u)ψ dx
Ω
≤ kψkLq(x) (Ω) k(|uj |q(x)−2 uj − |u|q(x)−2 u)kLq0 (x) (Ω) .
Therefore,
k(|uj |q(x)−2 uj − |u|q(x)−2 u)k(W 1,p(x) (Ω))0
0
Z
q(x)−2
= sup (|uj | uj − |u|q(x)−2 u)ψ dx
1,p(x) Ω
ψ∈W0 (Ω)
kψk 1,p(x) =1
W0 (Ω)
Proof of Theorem 1.3. In view of the previous result, we seek for critical values
below level c. For that purpose, we want to use the Mountain Pass Theorem.
Hence we have to check the following condition:
(1) There exist constants R, r > 0 such that when kukW 1,p(x) (Ω) = R, then
F(u) > r.
(2) There exist v0 ∈ W 1,p(x) (Ω) such that F(v0 ) < r.
Let us first check (1). We suppose that k|∇u|kLp(x) (Ω) ≤ 1 and kukLp(x) (Ω) ≤ 1.
The other cases can be treated similarly.
By Poincaré inequality (Proposition 3.1), we have
|∇u|p(x) |u|q(x) |u|r(x)
Z
− − λ(x) dx
Ω p(x) q(x) r(x)
kλk∞
Z Z Z
1 1
≥ |∇u|p(x) dx − |u|q(x) dx − |u|r(x) dx
p+ Ω q− Ω r− Ω
1 1 kλk∞
≥ k|∇u|kp+ − kukq−
Lq(x) (Ω)
− kukr−
Lr(x) (Ω)
p+ q− r−
1 C Ckλk∞
≥ k|∇u|kp+ − k|∇u|kq−L p(x) (Ω) − k|∇u|kr− Lp(x) (Ω)
.
p+ q− r−
1 p+
Let g(t) = p+ t − q− t − Ckλk
C q− ∞ r−
r− t , then it is easy to check that g(R) > r for
some R, r > 0. This proves (1).
1,p(x)
Now (2) is immediate as for a fixed w ∈ W0 (Ω) we have
lim F(tw) = −∞.
t→∞
Now the candidate for critical value according to the Mountain Pass Theorem is
c = inf sup F(g(t)),
g∈C t∈[0,1]
1,p(x)
where C = {g : [0, 1] → W0 (Ω) : g continuous and g(0) = 0, g(1) = v0 }.
1
We will show that, if inf x∈Aδ λ(x) is big enough for some δ > 0 then c < p+ −
1
n
q−
S and so the local Palais-Smale condition (Theorem 4.2) can be applied. We
A
1,p(x)
fix w ∈ W0 (Ω). Then, if t < 1, we have
|∇w|p(x) |w|q(x) |w|r(x)
Z
F(tw) ≤ tp(x) − tq(x) − λ(x)tr(x) dx
Ω p− q+ r+
tp− tr+
Z Z
≤ |∇w|p(x) dx − λ(x)|w|r(x) dx
p− Ω r+ Ω
tp− tr+
Z Z
≤ |∇w|p(x) dx − λ(x)|w|r(x) dx
p− Ω r+ Aδ
tp− tr+
Z Z
≤ |∇w|p(x) dx − ( inf λ(x))|w|r(x) dx
p− Ω r+ Aδ x∈Aδ
p− r+
We define g(t) := tp− a1 − (inf x∈Aδ λ(x)) tr+ a3 , where a1 and a2 are given by a1 =
k|∇w|p(x) kL1 (Ω) and a3 = k|w|r(x) kL1 (Aδ ) .
14 J. FERNÁNDEZ BONDER, A. SILVA EJDE-2010/141
1
The maximum of g is attained at tλ = (inf x∈Aa1λ(x))a3 r+−p− . So, we conclude
δ
that there exists λ0 > 0 such that if (inf x∈Aδ λ(x)) ≥ λ0 then
1 1
F(tw) < − − Sn
p+ qA
This completes the proof.
Remark 4.3. Observe that if λ(x) is continuous it suffices to assume that λ(x) is
large in the criticality set A.
4.2. Proof of Theorem 1.2. Now it remains to prove Theorem 1.2. So we begin
by checking the Palais-Smale condition for this case.
1,p(x)
Lemma 4.4. Let {uj }j∈N ⊂ W0 (Ω) be a Palais-Smale sequence for F then
{uj }j∈N is bounded.
1,p(x)
Proof. Let {uj }j∈N ⊂ W0 (Ω) be a Palais-Smale sequence; that is, F(uj ) → c
and F 0 (uj ) → 0. Therefore there exists a sequence εj → 0 such that
1,p(x)
|F 0 (uj )w| ≤ εj kwkW 1,p(x) (Ω) for all w ∈ W0 (Ω).
0
Now we have
1 0 1
c + 1 ≥ F(uj ) − F (uj )uj + − F 0 (uj )uj
q− q
Z Z
1 1 λ(x) λ(x) 1
≥ +− − |∇uj |p(x) dx + −
− − |uj |r(x) dx + − F 0 (uj )uj
p q Ω Ω q r q
We can assume that k|∇uj |kLp(x) (Ω) > 1. Then we have, by Proposition 2.5 and by
Poincaré inequality,
1 1 − 1 1 +
c + 1 ≥ + − − k|∇uj |kpLp(x) (Ω) + kλk∞ − − − kuj krLr(x) (Ω)
p q q r
1
− − kuj kW 1,p(x) (Ω) εj
q 0
1 1 − 1 1 +
≥ + − − k|∇uj |kpLp(x) (Ω) + kλk∞ − − − Ck|∇uj |krLp(x) (Ω)
p q q r
1
− − kuj kW 1,p(x) (Ω)
q 0
from where it follows that kuj kW 1,p(x) (Ω) is bounded (recall that p+ ≤ q − and
0
r+ < p− ).
Let {uj }j∈N be a Palais-Smale sequence for F. Therefore, by the previous
1,p(x)
Lemma, it follows that {uj }j∈N is bounded in W0 (Ω).
Then, by Theorem 1.1 we can assume that there exist two measures µ, ν and a
1,p(x)
function u ∈ W0 (Ω) such that
1,p(x)
uj * u weakly in W0 (Ω), (4.7)
|∇uj |p(x) * µ weakly in the sense of measures, (4.8)
q(x)
|uj | *ν weakly in the sense of measures, (4.9)
X
ν = |u|q(x) + νi δxi , (4.10)
i∈I
EJDE-2010/141 CONCENTRATION-COMPACTNESS PRINCIPLE 15
X
µ ≥ |∇u|p(x) + µi δxi , (4.11)
i∈I
1/p∗ (xi ) 1/p(xi )
Sνi ≤ µi . (4.12)
As before, assume that I 6= ∅. Now the proof follows exactly as in the previous
case, until we get to
Z Z
1 1 q(x) 1 1 n 1 1
c≥ + − − |u| dx + + − − S + kλkL∞ (Ω) + − − |u|r(x) dx.
p q Ω p q p r Ω
Applying now Hölder inequality, we find
Z
1 1 1 1
c≥ + − − |u|q(x) dx + + − − S n
p q Ω p q
1 1 q+
+ kλkL∞ (Ω) + − − k|u|r(x) kLq(x)/r(x) (Ω) |Ω| q− −r+ .
p r
If k|u|r(x) kLq(x)/r(x) (Ω) ≥ 1, we have
(q/r)−
c ≥ c1 k|u|r(x) kLq(x)/r(x) (Ω) + c3 − kλkL∞ (Ω) c2 k|u|r(x) kLq(x)/r(x) (Ω) ,
−
so, if f1 (x) := c1 x(q/r) − kλkL∞ (Ω) c2 x, this function reaches its absolute minimum
kλk ∞ (Ω) c2 (q/r)1− −1
at x0 = c1L(q/r) − .
On the other hand, if k|u|r(x) kLq(x)/r(x) (Ω) < 1, then
(q/r)+
c ≥ c1 k|u|r(x) kLq(x)/r(x) (Ω) + c3 − kλkL∞ (Ω) c2 kukLq(x)/r(x) (Ω) ,
+
so, if f2 (x) = c1 x(q/r) − kλkL∞ (Ω) c2 x, this function reaches its absolute minimum
kλk ∞ (Ω) c2 (q/r)1+ −1
at x0 = c1L(q/r) + . Then
(q/r) (q/r) − +
1 1 − −1 + −1
c≥ − − S n + K min{kλkL(q/r)
∞ (Ω) , kλkL(q/r)
∞ (Ω) },
p+ q
which contradicts our hypothesis. Therefore I = ∅ and so uj → u strongly in
Lq(x) (Ω).
With these preliminaries the Palais-Smale condition can now be easily checked.
1,p(x)
Lemma 4.5. Let (uj ) ⊂ W0 (Ω) be a Palais-Smale sequence for F, with energy
level c. There exists a constant K depending only on p, q, r and Ω such that, if
(q/r)− (q/r)+
1 − −1 (q/r)+ −1
− q1− S n +K min{kλkL(q/r)
c < p+ ∞ (Ω) , kλkL∞ (Ω) }, then there exists a subsequence
1,p(x)
{ujk }k∈N ⊂ {uj }j∈N that converges strongly in W0 (Ω).
The proof of the above lemma follows by the continuity of the solution operator
as in Theorem 4.2.
Assume now that k|∇u|kLp(x) (Ω) ≤ 1. Then, applying Poincaré inequality, we
have
1 + 1 − kλkL∞ (Ω) −
F(u) ≥ + k|∇u|kpLp(x) (Ω) − − kukqLq(x) (Ω) − −
kukrLr(x) (Ω)
p q r
1 + C − kλk L∞ (Ω) C −
≥ + k|∇u|kpLp(x) (Ω) − − k|∇u|kqLp(x) (Ω) − −
k|∇u|krLp(x) (Ω)
p q r
=: J1 (k|∇u|kLp(x) (Ω) ),
16 J. FERNÁNDEZ BONDER, A. SILVA EJDE-2010/141
+ − kλk ∞ C −
where J1 (x) = p1+ xp − qC− xq − L (Ω)
r− xr . We recall that p+ ≤ q − and
r− < r+ < p− < p+ .
As J1 attains a local, but not a global, minimum (J1 is not bounded below),
we have to perform some sort of truncation. To this end let x0 , x1 be such that
m < x0 < M < x1 where m is the local minimum and M is the local maximum
of J1 and J1 (x1 ) > J1 (m). For these values x0 and x1 we can choose a smooth
function τ1 (x) such that τ1 (x) = 1 if x ≤ x0 , τ1 (x) = 0 if x ≥ x1 and 0 ≤ τ1 (x) ≤ 1.
If k|∇u|kLp(x) (Ω) > 1, we argue similarly and obtain
1 − C + kλkL∞ (Ω) C +
F(u) ≥ k|∇u|kpLp(x) (Ω) − − k|∇u|kqLp(x) (Ω) − k|∇u|krLp(x) (Ω)
p + q r−
=: J2 (k|∇u|kLp(x) (Ω) )
where
1 p− C + kλkL∞ (Ω) C r+
J2 (x) = x − − xq − x .
p + q r−
As in the previous case, J2 attains a local but not a global minimum. So let x0 , x1
be such that m < x0 < M < x1 where m is the local minimum of j and M is
the local maximum of J2 and J2 (x1 ) > J2 (m). For these values x0 and x1 we can
choose a smooth function τ2 (x) with the same properties as τ1 . Finally, we define
(
τ1 (x) if x ≤ 1
τ (x) =
τ2 (x) if x > 1.
Next, let ϕ(u) = τ (k|∇u|kLp(x) (Ω) ) and define the truncated functional as follows,
|∇u|p(x) |u|q(x)
Z Z Z
λ(x) r(x)
F̃(u) = dx − ϕ(u) dx − |u| dx
Ω p(x) Ω q(x) Ω r(x)
Lemma 4.6. F̃ is C 1 , if F̃(u) ≤ 0 then kukW 1,p(x) (Ω) < x0 and F(v) = F̃(v) for
0
every v close enough to u. Moreover there exists λ1 > 0 such that if 0 < kλkL∞ (Ω) <
λ1 then F̃ satisfies a local Palais-Smale condition for c ≤ 0.
Proof. We have to check only the local Palais-Smale condition. Observe that every
Palais-Smale sequence for F̃ with energy level c ≤ 0 must be bounded, therefore
by Lemma 4.5 if λ verifies
(q/r)− (q/r)+
1 1 − −1 + −1
0< − − S n + K min{kλkL(q/r)
∞ (Ω) , kλkL(q/r)
∞ (Ω) },
p+ q
then there exists a convergent subsequence.
The following Lemma gives the final ingredients needed in the proof.
Lemma 4.7. For every n ∈ N there exists ε > 0 such that
γ(F̃ −ε ) ≥ n
1,p(x)
where F̃ −ε = {u ∈ W0 (Ω) : F̃(u) ≤ −ε} and γ is the Krasnoselskii genus.
EJDE-2010/141 CONCENTRATION-COMPACTNESS PRINCIPLE 17
1,p(x)
Proof. Let En ⊂ W0 (Ω) be a n-dimensional subspace. Hence we have, for
u ∈ En such that kukW 1,p(x) (Ω) = 1,
0
|∇(tu)|p(x) |tu|q(x)
Z Z Z
λ(x)
F̃(tu) = dx − ϕ(tu) dx − |tu|r(x) dx
Ω p(x) Ω q(x) Ω r(x)
|∇(tu)|p(x) |tu|q(x)
Z Z Z
λ(x)
≤ dx − ϕ(tu) dx − |tu|r(x) dx.
Ω p− Ω q+ Ω r+
If t < 1, then
Z p− Z q+ q(x)
t |∇u|p(x) t |u|
Z
inf x∈Ω λ(x) r+ r(x)
F̃(tu) ≤ −
dx − +
dx − t |u| dx
Ω p Ω q Ω r+
− + +
tp tq tr
≤ −
− + an − inf λ(x) + bn ,
p q x∈Ω r
where
nZ o
an = inf |u|q(x) dx : u ∈ En , kukW 1,p(x) (Ω) = 1 ,
0
Ω
nZ o
bn = inf |u|r(x) dx : u ∈ En , kukW 1,p(x) (Ω) = 1 .
0
Ω
Then
− + + − +
tp tq tr tp tr
F̃(tu) ≤ − an − inf λ(x) b n ≤ − inf λ(x) bn .
p− q+ x∈Ω r+ p− x∈Ω r+
Observe that an > 0 and bn > 0 because En is finite dimensional. As r+ < p− and
t < 1 we obtain that there exists positive constants ρ and ε such that
F̃(ρu) < −ε for u ∈ En , kukW 1,p(x) (Ω) = 1.
0
References
[1] Claudianor O. Alves; Existence of positive solutions for a problem with lack of compactness
involving the p-Laplacian. Nonlinear Anal., 51(7):1187–1206, 2002.
[2] Claudianor O. Alves and Yanheng Ding; Existence, multiplicity and concentration of positive
solutions for a class of quasilinear problems. Topol. Methods Nonlinear Anal., 29(2):265–278,
2007.
[3] Abbas Bahri and Pierre-Louis Lions; On the existence of a positive solution of semilinear ellip-
tic equations in unbounded domains. Ann. Inst. H. Poincaré Anal. Non Linéaire, 14(3):365–
413, 1997.
[4] Julián Fernández Bonder, Sandra Martı́nez, and Julio D. Rossi; Existence results for gra-
dient elliptic systems with nonlinear boundary conditions. NoDEA Nonlinear Differential
Equations Appl., 14(1-2):153–179, 2007.
[5] Alberto Cabada and Rodrigo L. Pouso; Existence theory for functional p-Laplacian equations
with variable exponents. Nonlinear Anal., 52(2):557–572, 2003.
[6] Teodora-Liliana Dinu; Nonlinear eigenvalue problems in Sobolev spaces with variable expo-
nent. J. Funct. Spaces Appl., 4(3):225–242, 2006.
[7] Pavel Drábek and Yin Xi Huang; Multiplicity of positive solutions for some quasilinear elliptic
equation in RN with critical Sobolev exponent. J. Differential Equations, 140(1):106–132,
1997.
[8] Xian-Ling Fan and Qi-Hu Zhang; Existence of solutions for p(x)-Laplacian Dirichlet problem.
Nonlinear Anal., 52(8):1843–1852, 2003.
[9] Xianling Fan and Dun Zhao; On the spaces Lp(x) (Ω) and W m,p(x) (Ω). J. Math. Anal. Appl.,
263(2):424–446, 2001.
[10] Yongqiang Fu; The principle of concentration compactness in Lp(x) (Ω) spaces and its appli-
cation. Nonlinear Anal., 71(5-6):1876–1892, 2009.
[11] J. Garcı́a Azorero and I. Peral Alonso; Multiplicity of solutions for elliptic problems with
critical exponent or with a nonsymmetric term. Trans. Amer. Math. Soc., 323(2):877–895,
1991.
[12] P.-L. Lions; The concentration-compactness principle in the calculus of variations. The limit
case. I. Rev. Mat. Iberoamericana, 1(1):145–201, 1985.
[13] Mihai Mihăilescu; Elliptic problems in variable exponent spaces. Bull. Austral. Math. Soc.,
74(2):197–206, 2006.
[14] Mihai Mihăilescu and Vicenţiu Rădulescu; On a nonhomogeneous quasilinear eigenvalue prob-
lem in Sobolev spaces with variable exponent. Proc. Amer. Math. Soc., 135(9):2929–2937
(electronic), 2007.
Analı́a Silva
Departamento de Matemática, FCEyN, Universidad de Buenos Aires, Pabellón I, Ciudad
Universitaria (1428), Buenos Aires, Argentina
E-mail address: asilva@dm.uba.ar