Академический Документы
Профессиональный Документы
Культура Документы
curvature operator
Alexandru Kristály 12
University of Babeş-Bolyai, Faculty of Economics, Cluj-Napoca, Romania
Email address: alexandrukristaly@yahoo.com
Csaba Varga2
University of Babeş-Bolyai, Faculty of Mathematics and Computer Sciences,
Cluj-Napoca, Romania
Email address: csvarga@cs.ubbcluj.ro
( p−2
Remark 0.1
i) Let p ≥ 2. If we consider K(s) = p1 [(1 + s2 ) 2 − 1], the we have K 0 (s) = k(s) =
p
p−2
s(1 + s2 ) p and we get the generalized mean curvature operator
p−2
div((1 + |∇u|2 ) 2 ∇u).
ii) We observe that the function k and K satisfies the following growth conditions:
and
1 p
|s| ≤ K(s) ≤ b(|s| + |s|p )
p
p
p−2
2 2 −1
for every s ∈ R, where a = 2 2 and b = p
.
(f1 ) f (0) = 0;
f (s)
(f2 ) lim p−1 = 0;
|s|→+∞ |s|
1
Supported partially by a Domus Hungarica Fellowship
2
Supported partially by the Research Center of Sapientia Foundation
1
Z s
(f3 ) Denoting by F (s) = f (t)dt, we also assume
0
Z Z
We denote by K(u) = K(∇u)dx and F(u) = F (u)dx.
Ω Ω
We introduce the functional Eλ : W01,p (Ω) :→ R, defined by Eλ (u) = K(u) −
λF(u). It is easy to see that the critical points of the functional Eλ are solutions of
problem (Pλ ).
Remark 0.2 The functional K is locally uniformly convex and K0 : W01,p (Ω) →
W0−1,p (Ω) verifies the (S+ ) condition, that is for every sequence (un ) ⊂ W01,p (Ω)
such that un * u and lim suphK0 (un ), un − ui ≤ 0 we have un → u strongly, see
n→+∞
Remark 3.1 and Proposition 2.1 [NM].
Proof. Because the functional K is locally uniformly convex follows that weakly
lower semicontinuous, seeRemark 0.2.
From the condition (f2 ) follows the existence of a number c ∈ R such that
Because the embeddings W01,p (Ω) ,→ Lp (Ω) is compact we obtain the sequentially
continuity of the function F.
Lemma 0.2 For every λ ∈ R \ {0}, the functional Eλ is coercive and satisfies the
Palais-Smale condition.
Proof. Let us fix λ ∈ R \ {0} arbitrarly. By (f2 ) there exists δ = δ(λ) such that
|f (s)| ≤ c−p −1
p (1 + |λ|) |s|
p−1
for every |s| ≥ δ, where cp denote the embedding constant of W01,p (Ω) ,→ Lp (Ω).
After integration of the above relation we have
1
|F (s)| ≤ c−p (1 + |λ|)−1 |s|p + max |f (t)||s|,
p p |t|≤δ
for all s ∈ R.
Thus for every u ∈ W01,p (Ω) we have
1 |λ|
Eλ (u) ≥ K(u) − |λ||F(u)| ≥ ||u||p − ||u||p
p p(1 + |λ|)
2
1 1 1
−|λ||Ω| q max cp ||u|| ≥ ||u||p − |λ||Ω| q max cp ||u||,
|t|≤δ p(1 + |λ|) |t|≤δ
where |Ω| denote the measure of Ω. If ||u|| → +∞, we have Eλ (u) → +∞, as well,
i.e. Eλ is coercive.
Now, let {un } ∈ W01,p (Ω) be a sequence such that {Eλ (un )} is bounded and
||Eλ0 (un )||W −1,p → 0. Since Eλ is coercive follows that, the sequence {un } is bounded
0
in W01,p (Ω). Up to a subsequence un * u weakly in W01,p (Ω) and un → u strongly in
Lp (Ω), because the embedding W01,p (Ω) ,→ Lp (Ω) is compact. From Eλ = K − λF
follows that
Z
0 0
hK (un ), un − ui = hEλ (un ), un − ui + λ f (un )(un (x) − u(x))dx. (0.1)
Ω
Because ||Eλ0 (un )||W −1,p → 0 and |hEλ0 , un − ui| ≤ ||Eλ0 (un )||W −1,p ||un − u|| follows
0 0
f (s)
that hEλ0 (un ), un − ui → 0. Because lim p−1 = 0, follows that the existence of
|s|→+∞ |s|
the number b1 , b2 > 0 such that
|f (s)| ≤ b1 + b2 |s|p−1 ,
for every s ∈ R.
Therefore
Z Z Z
|f (un (x))||un (x)−u(x)|dx ≤ b1 |un (x)−u(x)|dx+b2 |un (x))|p−1 |un (x)−u(x)|dx
Ω Ω Ω
1
≤ b1 |Ω)| q + b2 ||un ||p ||un − u||p .
Because the embedding W01,p (Ω) ,→ Lp (Ω) is compact follows that
Z
lim |f (un (x))||un (x) − u(x)|dx = 0.
n→+∞ Ω
Because the operator K0 has the (S+ ) property follows that un → u in W01,p (Ω).
Theorem 0.1 (see [Bo, Theorem 2.1]) Let X be a separable and reflexive real Ba-
nach space, and let K, F : X → R be two continuously Gâteaux differentiable func-
tionals. Assume that there exists x0 ∈ X such that K(x0 ) = F(x0 ) = 0 and K(x) ≥ 0
for every x ∈ X and that there exists x1 ∈ X, ρ > 0 such that
3
F(x1 )
(ii) sup F(x) < ρ .
K(x)<ρ K(x1 )
Further, put
ζρ
a= ,
ρ FK(x
(x1 )
1)
− supK(x)<ρ F(x)
with ζ > 1, assume that the functional K − λF is sequentially weakly lower semi-
continuous, satisfies the Palais-Smale condition and
Proof. Due to (f1 ), for an arbitrarly small ε > 0, there exists δ = δ(ε) > 0 such
that
|f (s)| ≤ εc−1
p for all s ∈ R.
Combining the above inequakity with
and 1
Sρ2 = {u ∈ W01,p (Ω) : bcp max{|Ω| q ||u||, cp−1 p
p ||u|| } < ρ}.
4
ρ
If ρ > 0 is sufficiently small and u ∈ Sρ2 , then ||u|| ≤ 1 and using (0.3) we
bcp |Ω| q
get
sup F(u) sup F(u)
u∈Sρ1 u∈Sρ2 ε 1 + δ ρp−1
0≤ ≤ ≤ + c p−1 p .
ρ ρ bcp δ |Ω| q
Because ε > 0 is arbitrarly and ρ → 0+ we get the desired result.
Our main result is the following.
References
[Bo] G. Bonanno, Some remarks on a three critical points theorem, Nonlinear Anal-
ysis TMA 54 (2003), 651-665.
[Br] H. Brézis, Analyse Fonctionnelle-Théorie et Applications, Masson, Paris, 1992.
[LY] G. Li and S. Yan, Eigenvalue problems for quasilinear elliptic equations on RN ,
Comm. Partial Differential Equations, 14 (1989), 1291-1314.
[MM] S. Marano and D. Motreanu, Existence of two nontrivial solutions for a class
of elliptic eigenvalue problems, Arch. Math. 75 (2000), 53-58.
[NM] P.D. Nápoli, M.C. Mariani, Mountain pass solutions to equations of p-
Laplacian type, Noonlinear Analysis, 54(2003), 1205-1219.
[Ra] P. Rabinowitz, Minimax Methods in Critical Point Theory with Applications
to Differential Eqautions, CBMS Reg. Conf. Ser. in Math. , Amer. Math. Soc.,
Providence, 1986.
[Ri1] B. Ricceri, On a three critical points theorem, Arch. Math. 75 (2000), 220-
226.
[Ri2] B. Ricceri, Existence of three solutions for a class of elliptic eigenvalue prob-
lems, Math. Comput. Modelling 32 (2000), 1485-1494.
[Sc] M. Schechter, Linking Methods in Critical Point Theory, Birkhäuser, Boston,
1999.
[Str1] M. Struwe, Variational Methods, Springer-Verlag, Berlin 1990.
[Str2] M. Struwe, Plateau’s problem and the calculus of variations, Math. Notes 35,
Princeton University Press, Princeton 1989.
5
[W] M. Willem, Minimax Theorems, Birkhäuser, 1995.