• Nie Znaleziono Wyników

Multiplicity of positive solutions for eigenvalue problems of (p,2)-equations

N/A
N/A
Protected

Academic year: 2022

Share "Multiplicity of positive solutions for eigenvalue problems of (p,2)-equations"

Copied!
17
0
0

Pełen tekst

(1)

Abstract

We consider a nonlinear parametric equation driven by the sum of a p-Laplacian

(p > 2) and a Laplacian (a (p,2)-equation) with a Caratheodory reaction, which is strictly (p - 2)-sublinea r near +oo. Using variational methods coupied with truncation

and comparison techniques, we prove a bifurcation-type theorem for the non lin ear eigenvalue problem. So, we show that there is a critical parameter value A,> Osuch that for A> A• the problem has at leasttwo positive solutions, if A= A•, then the problem has at least one positive solution and for A

E

(0,A,), it has no positive solutions.

MSC: 35J25; 35J92

Keywords: nonlinear regularity; tangency principle; p-Laplacian; bifurcation-type theorem; positive solutions

Gasiriski and Papageorgiou Boundary Value Problems 2012,

2012:152

http://www.bounda ryval ueproblems.com/content/2012/1/152 v> Boundary Value Problems

a SpringerOpen Journal

RES EARCH OpenAccess

Multiplicity of positive solutions foreigenvalue problems of (p, 2)-equations

Leszek GasinskiF and Nikolaos S. Papageorgiou

2

'Correspondence:

Lesze k.Gasinski@ii.uJedu. pi

1Faculty of Mathematics and Computer Science, Institute of Computer Science, Jagiellonian University, ul.lopsiewicza 6, Krakow, 30-348, Poland Full list of author information is available at the end of the article

1 Introduction

Let Q<;;;RN be a bounded domain with a C

2

-boundary BQ. In this paper, we study the following nonlinear Dirichlet eigenvalue problem:

! - A pu(z)-Au(z) = )f(z,u(z)) in

Q,

ulan=O, u>O, A>O, 2<p<+oo. (Ph

Here, by Ap we denote the p-Laplace differential operator defined by

Springer

(with 2 <p <+oo). ln (Ph, >.. >0 is a parameter andf(z, t) is a Caratheodory function (i.e., for all t

E

R, the function z

t----+

f(z,t) is measurable and for almost all z

EQ,

the func tion t

t----+

f(z,t) is continuous),which exhibits strictly(p- 2)-sublinear growth int h e

t-variable near +oo. The aim of this paper is to determine the precise dependence of the set of positive solutions on the parameter>.. > 0. So, we prove a bifurcation-type theorem, which establishes the existence of a critical parameter value A•> 0 such that for all). >).,, problem (Ph has at leasttwo nontrivial positive smooth solutions, for>..=>..,, problem (Ph has at least one nontrivial positive smooth solution and for).

E

(O,>..,), problem (Ph has no positive solution. Similar nonlinear eigenvalue problems with(p- 2)- sublinear reac tion were studied by Maya and Shivaji [l] and Rabinowitz [2] for problems driven by the Laplacian and by Guo [3], Hu and Papageorgiou [4] and Perera [5] for problemsdriven

©2012 Gasinski and Pa pageorgiw;licensee Springer. This is an Open Access article distr iru t ed under the terms of the Creative Commons Attribution License (http//creativecommons.org/licenses/by/2.0), which permits unrestricted use, di st r ibuti on,and re production in any medium, provided the original work is pro per lycited.

(2)

Page

2

of17 Gasir'iskiandPapageorgiouBoundaryValueProblems2012,2012:152http://www

.boundaryvalueproblems.com/content/2012/1/152

bythep-Laplacian.However,noneoftheaforementionedworksproducestheprecisede

pendenceofthesetofpositivesolutionsontheparameter J.. >0 (i.e.,theydonotprove a bifurcation-

type theorem). We mention that in problem(Phthe

differentialoperatorisnothomogeneousincontrasttothecaseoftheLaplacianandp- Laplacian.Thisfactis

thesourceofdifficultiesinthestudyofproblem(Phw h i c h leadtonewtoolsandmeth ods.

We point out that (p, 2)-equations (i.e., equations in which the differential operatoristhesum of ap-Laplacian and a Laplacian) are important in quantum physics in

the search for solitions.WerefertotheworkofBenci,D'Avenia-

FortunatoandPisani[6].Morerecently, there have been some existence and multiplicity results for such problems; see Cingolani and Degiovanni [7], Sun [8]. Finally, we should mention the recent papers of Marano and Papageorgiou [9, 10]. In [9] theauthorsdeal with parametric p-Laplacian equations in which the reaction exhibits competing nonlinearities(concave-convexnonlinearity). In [10],theystudya nonparametric (p, q)- equation with a reaction that hasdifferent behavior bothat±oo and at O from those considered in the present paper,andso the geometry of the problem isd i f fe re nt .

Outapproachisvariationalbasedonthecriticalpointtheory,combinedwithsuitable truncation andcomparison techniques. In the next section, for theconvenienceof the reader,webrieflyrecallthemainmathematicaltoolsthatweuseinthispaper.

2 Mathematicalbackground

LetXbe a Banach space and let x' be its topological dual. By(·,·)we denote the dual ity brackets for the pair (X' ,X).Let cp

E

C

1

(X). A point x

0E

X is a critical point of cp if cp'(x

0 )

=0.Anumberc EE isa criticalvalue of cp ifthereexistsacriticalpoint x

0

EX s u ch that cp(xo) = c.

Wesaythat cp

E

C

1

(X) satisfiesthePalais-

Smale conditionifthefollowingistrue:'Everysequence {xnln:0:1 c:; X, suchthat{ cp(xn) ln:0:1 c:;Eis

boundedand

admits a strongly convergent subsequence:

Thiscompactness-typeconditioniscrucialinprovingadeformationtheoremwhichin turnleadstotheminimaxtheoryofcertaincriticalvaluesof cp

E

C

1

(X)(see, e.g., Gasinski andPapageorgiou[11]).Awell-writtendiscussionofthiscompactnessconditionandits roleincriticalpointtheorycanbefoundinMawhinandWillem[12].Oneoftheminimax theoremsneededinthesequelisthewell-known'mountainpasstheorem'.

Theorem2.1 Jf cp E C

1

(X ) satisfies the Palais-Smalecondition,x

0

,x

1

EX,[[x

1

-x

0

II > r >0,

max{cp(xo),cp(xi)} < inf{cp(x):[!x - xoll

=

r}

=

T/r

and

c = inf max cp(y(t)),

y

Er

O::t::l

(3)

with a

0E

L

00

( Q)+, c

0

> 0and1< r <p',where

J

where

r = {y

E

C((O,l];X):y(O) = Xo, y(l) =xi}, thenc T/rand cisacriticalvalueof cp.

Intheanalysisof problem(P),.,inadditionto the Sobolev space wt·P(Q), we willalsouse the Banach space

Thisisan ordered Banachspacewithapositivecone:

C+ ={u

E

C6(Q): u(z) 0 for all z

E

Q } . This conehasanonemptyinterior givenby

int C+ = { u E C+: u(z) >0for all z E Q, :: (z)<0 for all z E oQ},

where by n(·) we denote the outward unitnormalon oQ.

Letf 0:Q x R--+ R bea Caratheodoryfunction with subcriticalgrowthins E R, i.e.,

l.fo (z,t) I::'. ao(z) +cols

1r-l

for almost all z

EQ,

alls

E

R ,

p . = 1

p

ifp<N,

+ooi f p N

(the critical Sobolev exponent).We set

andconsider theC

1

-functional 1/to:wt·P(Q)--+ Rdefined by

1/to(u) = l!v'u[I + 11vu11- f Fo (z, u(z))dzVu

E

wt·P(Q).p 2 (2.1)

Thenext proposition isaspecialcaseofamoregeneral resultproved byGasinski andPapageorgiou [13]. We mention that thefirstresultofthis type was provedbyBrezisandNirenberg (14].

Proposition 2.2Jf1/tois defined by(2.1) andu

0E

W6'P(Q)isalocalC5(Q)-

minimizerof1/to,i.e.,thereexists

Qi

> 0 suchthat

(4)

thenu

0E

C,,6(0..)forsome j3

E

(O,1) andu

0

isalsoalocalWJ'P(0..)-minimizer of 1/ro, i.e.,there exists Qz > 0 suchthat

Let g,h

E

l

00

(0..).We say that g-<hif for all compact subsets

1(

r:::: 0..,wecanfinds = s(I() > 0suchthat

g(z) +

s

S h(z) for almostall z

E

K.

Clearly,ifg,h

E

C(0..)andg(z) < h(z) forall z

E

0..,t hen g-< h. A slight modification of the proof of Proposition 2.6ofArcoyaandRuiz [15) in order toaccommodatethe presence of the extra linear term -,1'),.u leads to the following strong comparisonp r i n c i p l e .

Proposition 2.3 If 2:': 0, g,h

E

l

00

(0..), g-<handu

E

C5(0..), v

E

intC+ aresolutions oftheproblems

-,1'),.pu(z)-,l'),.u(z) + [u(z)[P-

2

u(z) =g( z) {-,1'),.pv(z)-,l'),.v(z) + [v(z)[P-

2

v(z) = h(z)

in 0.., in 0.., then v- u

E

intC+.

Let r

E

(1, +oo) andlet Ar: wJ·r(0..)---+w-V (0..)

=

WJ'(0..)' (where +

=

1)bea nonlinear map definedb y

(2 .2 ) Thenext proposition can be found in Dinca, JebeleanandMawhin [16)andGasinskiandPapageorgiou [11).

Proposition 2.4 If Ar: WJ'(0..)---+ w-V (0..) (where 1< r < +oo) isdefined by (2.2), then Ar iscontinuous,strictly monotone (hence maximal monotone too),boundedand of type (S)+, i.e., ifun---+ u weaklyin WJ'(0..) and

lim su p(Ar(un), Un-u) S0,

n --++00

Ifr= 2,thenwewriteA

2

= A

E

£(HJ(0..);H-

1

(0..)).

Inwhatfollows,by 7:

1(

p)wedenotethefirsteigenvalueofthenegativeDirichletp- Laplacian (-,1'),.p, wt·P(0..)). We know that A

1

(p)>0anditadmitsthe following varia tional characterization:

(2.3)

(5)

Finally,throughoutthiswork,by I I·I Iwe denotethenormoftheSobolevspace wJ·P (Q).By virtue of the Poincare inequality, weh a v e

The notation II ·II will also be used to denote the norm ofRN.No confusion is possible since it willalwaysbe clear from the context which norm is used.

FortER,wesett±=max{±t ,O}.Then for u EwJ·P(Q), we define u±O = u(·)±. We know that

If h: QxR---+ Rissuperpositionallymeasurable (for example, a Caratheodory function), then we set

By I· INwe denote the Lebesgue measure onRN.

3 Positivesolutions

The hypotheses on the reaction!arethefollowing.

H: f:QxR---+RisaCaratheodory function such thatf(z, 0) = 0 for almost all z EQ, f(z,t) ?::. 0 for almost allzEQ and all t ?::.0 and

(i) forevery I?> 0, thereexists ae E l

00

(Q+)suchthat f(z, t) :'S ae(z) for almost all

Z

E Q, all t E [0, Q);

(ii) limt-++oo/= 0 uniformly for almost allzEQ;

(iii) limt-+o+ '.!/ = 0 u n if o r m ly foralmostall z EQ;

(iv) for every

Q

>0, there existse >0 such that for almostallzEQ, themap t

1---+

f(z, t) + Ptp-l is nondecreasing on [O,

Q);

(v) if

F(z, t) 1 = t f(z,s)ds,

then there exists c ERsuch that F(z, c) > 0 for almost all z EQ.

Remark 3.1Since we are looking for positive solutions and hypotheses Hconcernonly the

positive semiaxisR+= [O, +oo), we may and will assume thatf(z, t) = 0 for almost all

z EQ and all t -c: 0.Hypothesis H(ii) implies that foralmostall z EQ, the map f(z,·) is strictly(p-

2)-sublinear near +oo. Hypothesis H(iv) is much weaker than assuming the

monotonicity off(z,·) for almost all z EQ.

(6)

Page6of17 Gasir'iski and PapageorgiouBoundaryValue Problems2012,

2012:152htt p://w

ww.boundaryvalueproblems.com/contet/n2012/1/152

Example3.2 The followingfunctions satisfy hypotheses H (for the sake of simplicity, we dropthez-dependence):

with1< q < p <r< 1/. Clearly .h is not monotone. Let

Y ={ J..> 0: problem (Ph has a nontrivial positive solution}andlet S(J..) be the setofsolutionsof (Ph. We set

J.., = infY

(if Y

=

0, then J..,

=

+oo).

Proposition 3.3 If hypotheses H hold, then S(J..)

<::::

int C+ and ).,>0.

Proof Clearly,the result is true if Y =0.So,suppose that Yi 0andlet J..

E

Y. So,wecan find u

E

S(J..) n wt·P(Q)suchthat

-L">pu(z)-L">u(z)=J..f(z,u(z)) { u ian= 0.

in

Q,

FromLadyzhenskayaandUraltseva[17,p.286],wehavethat u

E

l°"Q(). Thenwecan applyTheorem1ofLieberman[18]a n dhavethat u

E

intC+\ {O).Let e = llulloo andlet te > 0beaspostulated by hypothesis H(iv). Then

-L">pu(z)-L">u(z) + teu(z)P-1 :"': 0 foralmost allz

EQ ,

so

L">pu(z)+!'.u(z)::Steu(z)P-l fora l m os t allz

EQ.

From thestrongmaximum principleofPucciand Serrin[19,p.34],we have that

u(z)> 0 Vz

EQ.

So,wecanapply the boundary point theoremofPucciandSerrin [19,p.120] andhavethat

u

E

intC+.Therefore, S(J..)

<::::

intC+.

(7)

2012:152http://www.boundaryvalueproblems.com/content/2012/1/15 Page 2

- -

Byvirtueofhypotheses H(ii)and(iii),we see that wecanfindc

1

> 0 such that

j(z, t): ' . ' :

C1tp-l

for almost all

ZEQ,

all t'.::': 0. (3.1)

LeU...o

E

(0,ir))andiJ

E

(0,J...o].SupposethatiJ

E

Y. Thenfromthefirstpartoftheproof,we know that wecan find u

6E

S( iJ) <:;; int C+. We have

so

(see(3.1)andrecallthatiJ :::J...o < ;:rl), whichcontradicts(2.3).Therefore,J...,'.::':J...o >0.

D ForJ...>0,let ({!J..:wt·P(Q)--+ be the energy functional for problem(Phdefinedby

Evidently, rpJ.. EC(WJ·P(Q)).

Proposition 3.4 If hypotheses H hold,then Y / 0.

Proof By virtueofhypotheses H(i)and (ii), foragiven s > 0,wecanfind c

6

> 0 such that F(z,t):'.': -tP +

s

C

8

foralmostall z

EQ,

all t'.::': 0.

p (3.2)

Thenfor u

E

wt·P(Q) andJ...>0,we have

({!J.. (u) = 1 - llv'ullf +1 -II V ul@- 1 J... F (z, u) dz

p 2 n

'.::': 1 11vu11rp - AS [[u+[[

p

- ACslQIN

p p

p

(3.3)

(see (3.2) and (2.3)).

Lets

E(0,

IiJ:l).Thenfrom(3.3)itfollowsthat ({!J..is coercive.Also,exploitingthecom

pactnessoftheembedding W6'P(Q)<:;;LP(Q) (bytheSobolevembeddingtheorem),weseethat (f!

J..is sequentiallyweaklylowersemicontinuous.So,bytheW e i e r s t r a s s theorem,wecanfind u

0E

wt·

P(Q) suchthat

({!J..(uo)= inf ({!J..(u).

UEW6'p(Q)

(3.4)

(8)

Considertheintegralfunctional K: LP(Q)---+ definedby K(u)= l F(z,u(z))dz VuELP(Q).

Hypothesis H(v) implies thatK(c)>0 and sinceF(z, t) = 0 for almost allz

E

Q,alls :s 0, we mayassumethat c >0.SincewJ·P(Q)isdenseinLP(Q)and c >0,wecanfind v

E

W t ' r ( Q ) ,

i> 2:0, such thatK(v)>0. Then for)._> 0 large,we have

so

cp;,,. (i>) < 0 for)._

> 0 largeandth u s cp;,,. (uo) <0 = cp;,,.(0)

(see (3.4)), hence u

0

f O. From (3.4), we have cp{ (uo) = 0,

so

(3.5) On(3.5),weactwith -Uo

E

wt,p(Q).Then

hence uo 2:0, uo f O.

From (3.5),we have

-Apuo(z) -Auo(z) =Aj(z, uo(z)) in0 . , { uo[an =0, uo 2:0, uo iO,

so u

0E

S()._) c:; int C (see Proposition 3.3).

So, for)._2:)._,big, we have)._

E

Y andso Yi 0 . D

Proposition 3.5 If hypothesesHhold and)._

E

Y, then[A, +oo) c:; Y.

Proof Since by hypothesis)._

E

Y , we canfind a solution u;,,.

E

int C of (Ph (seePropo sition 3.3).

Let µ,, >) . _ and co nsid er thefollowingtruncation of the reaction inproblem (P)

µ.:

hµ.(Z, t) = {µ,,f(z,u;,,.(z)) µ,,f(z, t)

if S'.': U;,,.(z),

ifu;,,.(z)< t. (3.6)

(9)

1

This is a Caratheodory function. Let

andconsidertheC 1 -functional 1/tµ,:w JP·(Q)----+' de finedb y

As in the proof of Proposition 3.4,using hypotheses H(i) and (ii),we see that 1/tµ, iscoercive.

Also, it is sequentially weakly lower semicontinuous. So, wecan find uµ,EwJ·P(Q) such that

so

andthus

(3.7) On(3.7) we act with (u,.-uµ,)+E wJ·P(Q). Then

(Ap(uµ), (u,.-uµ,)+) + (A(uµ),(u,.-uµ,)+)

= l hµ,(z,uµ,)(u,.-uµ,)+ dz

= l µf(z,u,_)(u,_-uµ)+dz 2: l ).j(z,u,.)(u,.-uµtdz

= (Ap(u,.),(u,.-uµ,)+) + (A(u,_),(u,.-u µ,t) (see(3.6) and use the facts that µ,>).andf2:0), so

(ll'vu,.I IP-

2

V

u,.-l l'vuµ,I-IP

2

V

uµ,, 'vu,.-'vuµ,)RNdz + IIV(u,. - uµ)+ II S0,

{u,,_>u;;)

thus

and hence u,. s uµ,.

T herefo re,(3.7) becomes

(10)

Page 10of17 PapageorgiouBoundaryValueProblems2012,2012:152http://www.boun

daryvalueproblems.com/content/2012/1/152

so

henceµ.

E

Y. Thisproves that [J.., +oo)<;; Y. D

Proposition3.6 If hypotheses H hold,then for every J.. >J..,problem (Phhasat least twopositive solutions

Uo,U

E

intC+, Uo /u.

Proof Note that Proposition 3.5impliesthat (J..,, +oo)<;; Y. Let J.., < rJ < J.. <µ..Then wecanfind UfJ ES(rJ)<;; intC+ and uµ, ES(µ.)<;; intC+. We have

-ApUfJ-Au0 rJf(z,UfJ) ::S Aj(z,UfJ) in

Q,

- Apuµ,-Auµ,µ.f(z,uµ,) 2: Aj(z,uµ,) in Q

(3.8) (3.9)

(recall thatf2:0and rJ < J.. <µ.).As inthe proofofProposition3.5,wecanshow that UfJ :s uµ,. Weintroducethe followingtruncationofthe reaction in problem (Ph:

Aj (Z,UfJ(z)) if t < UfJ(Z),

g,.(z,t) = { Aj(z,t) ifu0(z) :St ::S uµ,(z), (3.10)

Aj (z,uµ,(z))if uµ,(z)< t.

Thisis aCaratheodoryfunction. We set

andconsider theC

1

-functional f,.: Wt'P(Q)---+ Rdefined by

It is clear from(3.10) that f,. iscoercive.Also,it is sequentially weakly lower semicontin uous.So,wecan find u0

E

w ·P(Q) such that

so

andthus

(3.11)

(11)

PapageorgiouBoundaryValueProblems2012,2012:152http://www.boun

daryvalueproblems.com/content/2012/1/152 Page

11

of17

Acting on (3.11) with (u 0- uo)+

E

wt·P(Q) andnext with (u0 - u1 ,)+

E

wt·P(Q) (similarly as in theproofof Proposition 3.5), we get

Hence, we have

where [u

0,

uµ] ={ u

E

wt·P(Q):u

0

(z) :"Su(z):"Suµ (z) for almost all z

E

Q}.

Then (3.11) becomes

(see (3.10)), so

uo

E

S(A)

<;::;

intC+.

Let

a(y)= i!Ypii-

2

y + YVy

E

N.

Thena

E

C

1

(RN;RN)(recallthatp > 2)and

so

Note that

So, we can apply thetangencyprinciple of Pucci and Serrin [19, p.35] and inferthat

u 0 (z) < uo(z)Vz

EQ .

(3.12)

Let e = !!uo l loo and let(! > 0 be as postulated by hypothesis H(iv).Then

- !'.puo (z)- !'.uo (z)+ iJ(!u0 (z)P-

1

= iJf (z,uo(z)) + iJ(!uo(z )P-l

:"SiJf(z,uo(z)) + iJ(!uo(z)P-l

:"S Af(z,uo(z)) + iJ(!uo(z)P-l

= -!'.puo(z)- !'. u0(z) + iJ(!u0(z)P-l foralmostallz

E

Q

(12)

1

(see hypothesisH(iv) anduse the factsthan> iJ andf

:=::_

0), so

uo-uo

E

intC+ (3.13)

(see (3.12) and Proposition 2.3).

In a similar fashion,we show that

uµ,-u

0E

intC+. (3.14)

From (3.13)and(3.14), it follows that

(3.15) From (3.10),we see that

forsome;

E

R.

So,(3.15)impliesthat u

0

isalocalC6(Q)-minimizerof cp;._. InvokingProposition2.3,we havethat u

0

isalocal wJ·P(Q)-minimizerof cp;._.

(3.16)HypothesesH(i), (ii) and (iii) implythatforgiven s > 0 and r >p, wec an f i n d c2=c2(s,r) >

0 such that

F(z, n :s - c tp + c2tr foralmostallz

E

Q,all t

:=::.

0.

p (3.17)

Then for all u

E

W6'P (Q), we have

cp;._ (u)= 1 - llv'ull;+ 1 -II v'ul@-AF(z,u)dz

p 2 n

1 1 Ac

p

r

:=::.

- p llv'ullPP + - 2 llv'u l@-- p llu+II

p

-Ac2llu+II

r

(3.18)

for somec

3

>0 (see (3.17)and(2.3)).

Choose s

E

(0,iiJl:).Then,from(3.18)andsince r >p,weinferthat u

0

isalocalminimizer

of cp;._. Without any loss of generality, we may assume thatcp;._(O)= 0 :s cp;._(u

0)

(the analysis is similar if the opposite inequality holds). Byvirtueof (3.16), as in Gasinskiand

Papageorgiou [20] (see the proof of Theorem 2.12), wecanfindO < 12 < llu o II such

thatcp;._(O)

=

0:S cp;._(uo) < inf{cp;._(u): llu-uoll

=

12}

=

7/· (3.19)

(13)

1

Recall that cp,. iscoercive,hence it satisfies the Palais-Smalecondition.This fact and(3.19)permit the use of the mountain pass theorem(seeTheorem 2.1). So, we can findft

E

wt·P(Q)suchthat

(3.20) and

cp{ (u) = o. (3.21)

From (3.20) and(3.19),we have that u JO, u = u

0•

From (3.21), it follows that u

E

S(J..)

<:;::

- D

Next, weexaminewhat happens at thecriticalparameter J..,.

Proposition3.7IfhypothesesHhold, then).,

E

Y.

ProofLet P-.nl n:':l

<:;::

Y beasequencesuchthat

A• <An Vn::>:1

and

An A• asn----++oo.

Foreveryn ::>: l,we can find

UnE

int C+, such that

(3.22)

We claim that the sequence {unln:':l

<:;::

Wt'P(Q)is bounded. Arguing indirectly, suppose that the sequence

{u n)n:':l<:;::

Wt'P(Q)is unbounded. By passing to a suitable subsequence if necessary, we may assume that !!UnI I-+ + o o . L e t

Yn =

Un

l unll Vn>1.

Then l!YnI I = = 1and Yn

E

int C+foralln ::>: 1. From(3.22),wehave

Recall that

f(z, s)'.S

C1tp-l

foralmostall

ZE Q,

all t 2: 0

(3.23)

(see (3.1)),so the sequence{ 1 < ;l

1

ln:':l

<:;::

LP (Q) is bounded. Thisfact and hypothesis H(ii) imply that at least for a subsequence, we have

(3.24)

(14)

(seeGasinskiandPapageorgiou[20]).Also,passingtoasubsequence ifnecessary ,wemayassumet h a t

Yn---+ yweakly inwJ·P(Q), Yn---+Y in LP(Q).

(3.25) (3.26)

On (3.23) weactwith Yn- y E WJ'P(Q),pass to the limitas n---++oo anduse (3.24)and (3.26).Then

so

Using Proposition2.4, we have that

andso

IIYII = 1. (3.27)

Passingtothelimitas n---++ooin (3.23)andusing(3.24),(3.27)andthefactthatp >2, we obtain

soy= 0,whichcontradicts (3.27).

Thisprovesthatthesequence {unnlc>:l<;;;w·P(Q) isbounded.So,passingtoasubse- quence if necessary, we may assume that

Un---+U• weaklyin Wp' Un---+ U• inLP(Q).

(Q), (3.28)

(3.29)

On (3.22) weactwith Un-u,

E

w ·P(Q), pass tothelimitasn----,. +oo anduse(3.28) and(3.29).Then

lim((Ap(un),Un-U•)+(A(u n), Un-U•)) = 0,

n---++oo

so

limsu p(Ap(Un),Un-U•)'.':0

n ---++00

(15)

(since A is monotone) and thus

(see Proposition 2.4).

Therefore, if in (3.22) we pass to the limit asn----,.+ooand use(3.30),then

and so u,

E

C+ is a solution of problem(P),..,.

We need to show that u, fO. From(3.22),we have

(3.30)

- A pun(z )-Aun(z) = AJ (z,Un(z)) in Q

{Un Ian= 0 'in?::1.

From Ladyzhenskaya andUraltseva(17, p.286],we know that we can find M

1

> 0 such that

Then applying Theorem 1 of Lieberman (18], wecanfind fJ

E

(0,1) and M

2

>0 such that

Recall that C'/J(Q) is embeddedcompactlyin q(Q).So ,by virtue of (3.28), we have Un---+U• in CMQ).

Suppose that u, = 0. Then

(3.31)

Hypothesis H(iii) implies that foragiven s >0, we can find8

E

(0, s] such that f(z, 0 :'.S stp-l foralmostall z

EQ,

all t

E

(0,8].

From (3.31),it follows that wecanfind n

0

?:: 1 s uch that un(z)

E

(0,8] Vz

EQ,

all n?::no.

Therefore, for almost all z

E

Q and all n ?:: n

0,

we have

(see (3.32)and (3.33)), so

(3.32)

(3.33)

(16)

(see (2.3)),thus

and so

Lets \.0 to geta contradiction. This proves that u, /0 andsou, ES().,)<;;; int C, hence

).,E

Y. D

The bifurcation-type theorem summarizes the situation for problem(P h .

Theorem 3.8 If hypothesesHhold, then thereexists ).,>0such that (a) forevery).>).,problem(Phh a s atleasttwopositivesolutions:

(b) for).=).,problem(Phhas at least one positive solution u,

E

int C+;

(c) for).

E

(0,).,)problem(Phh a s nopositivesolution.

Remark 3.9Asthe referee pointed out, it isan interesting problem to produce an example in which, at the bifurcation point).' > 0, the equation has exactlyone solution. We believe that the recent paper of Gasinski and Papageorgiou [21] on the exist ence and uniqueness of positive solutions will be helpful. Concerning theexistence of nodal solutions for).

E

(0,).'),we mention the recent paper of Gasinski and Papageorgiou [22], which studies the (p, 2)-equations and produces nodal solutions for them.

Competing interests

The authors declarethattheyhaveno competinginterests.

Authors' contributions

Theauthorsdeclarethattheworkwasrealizedincollaborationwiththesameresponsibility.Allauthorsreadandapprovedthefinalmanu script.

Author details

1Facultyof MathematicsandComputer Science,Institute of ComputerScience,JagiellonianUniversity,ul.l ojasiewicza6,Krakow,30-348,Poland.' Department ofMathematics,National TechnicalUniversity,Zografou Campus,Athens,15780,Greece.

Acknowledgements

Dedicated to Professor JeanM awhin on the occasion of his70thbirthday.

Theauthorswouldliketo expresstheirgratitudeto bothknowledgeablereferees for theircorrections andremarks.Thisresearch hasbeen partiallysupportedbytheMinistryofScienceandHigher Edu cationofPolandunderGrantsno.N201542438 andN201604640.

Received: 13September 2012 Accepted: 7 December 2012 Published:28 December2012

References

1. Maya,C,Shivaji,R:Multiplepositives olutionsforaclassofsemilinearellipticboundaryvalueproblems.Nonlinear Anal.38,497- 504(1999)

2. Rabinowitz , PH:Pairsofpositivesolutionsofnonlinearellipticpartialdifferentialequations.IndianaUniv.Math.J23,173-186(1973) 3. Guo,Z:Someexistenceandmultiplicityresultsforaclassofquas ilinearellipticeigenvalue problems.NonlinearAnal. 18,957-

971(1992)

(17)

doi:1O.l 186/1687-2770-2012-152

Cite this article as: Gasi nski and Papageorgio u MuitipIicity of positive solutions for eigenvalue probi ems of (o, 2)-equations.

Boundary Value Problems

2012 2012152.

Submit your manuscript to a Springer0 pen° journal and benefit from:

.,. Convenient online submission .,. Rigorous peer review

.,. Immediate publication on acceptance .,. Open access: articles freely available online .,. High visibiilty within the field

.,. Retaining the copyright to your article

Submit your next manuscript at .,. sp ringeropen.com GasiriskiandPapageorgiou BoundaryValueProblems

2012,2012:152

http://www.bounda ryvalueproblems.com/content/2012/1/152

4. Hu, S, Papageorgiou, NS: Multiple positive solutions for nonlineareigenvalue problems with the p-Laplacian.

Nonlinear Anal.69,4286-4300(2008)

5. Perera, K: Multiple positive solutions of a class of quasilinear elliptic boundary value problems. Electron.J.Differ.E q u . 7,1-5 (2003)

6. Benci,1/,D'Avenia,P,Fortunato,D,Pisani,L:Solitonsinseveralspacedimensions:Derrick's problemandinfinitely many solutions.

Arch. Ration. Mech. Anal.154,297-324(2000)

7. Cingolani, S, Degiovanni, M: Nontrivial solutions for p-Laplace equations with right hand side having p-lineargrowth at infinity. Commun. Partial Differ. Equ. 30, 1191-1203(2005)

8. Sun, M:Multiplicity of solutions for a class of the quasilinear elliptic equations at resonance.J.Math. Anal. Appl.386, 661-668(2012)

9. Marano, SA, Papageorgiou, NS: Multiple solutions to a Dirichlet problem with p-Laplacian and nonlinearity depending on a parameter. Adv. Nonlinear Stud. 3, 257-275( 2 0 1 2 )

10. Marano, SA, Papageorgiou, NS: Constant-sign and nodal solutions of coercive (o,q)-Laplacianproblems.Nonlinear Anal.77,1 1 8 - 1 2 9 ( 2 0 1 3 )

11. Gasinski,L,Papageorgiou,NS:NonlinearAnalysis.Chapman&Hall/CRC,BocaRaton(2006)

12. Mawhin,J,Willem, M: Origin and evolution of the Palais-Smale condition in critical point theory.J.Fixed Point Theory Appl.7,265-290(201O)

13. Gasinski, L, Papageorgiou, NS:Multiple solutions for nonlinear coercive problems with a nonhomogeneous differential operator and a nonsmooth potential. Set-Valued Var. Anal. 20, 417-443(2012)

14. Brezis,H,Nirenberg,L:H1versusC1localminimizers.C.R.Acad.Sci.,Ser.1Math.317,465-472(1993)

15. Arcoya, D, Ruiz, D: The Ambrosetti-Prodi problem for the p-Laplace operator. Commun. Partial Differ. Equ. 31, 849-865 (2006)

16. Dinca, G, Jebelean, P, Mawhin,J:Variational and topological methods for Dirichlet problems with p-Laplacian. Port.

Math.58,339-378(2001)

17. Ladyzhenskaya, OA, Uraltseva, N: Linear and Quasilinear Elliptic Equations. Mathematics in Science and Engineering, vol. 46. Academic Press, New York(l968)

18. Lieberman, GM: Boundary regularity for solutions of degenerate elliptic equ at ions.Nonlinear Anal.12,1 2 0 3 - 1 2 1 9 (1988)

19. Pucci,P,Serrin,J:TheMaximumPrinciple.Birkhauser,Basel(2007)

20. Gasinski, L, Papageorgiou, NS: Nodal and multiple constant sign solutions for resonantp-Laplacian equations with a nonsmooth potential. Nonlinear Anal.71,5747-5772(2009)

21. Gasinski, L, Papageorgiou, NS: Existence and uniqueness of positive solutions for the Neumann p-Laplacian. Positivity (2012).doi:10.1007/s11117-012-0168-6(publishedonline)

22. Gasinski, L, Papageorgiou, NS: Multiplicity theorems for (o,2)-equations(submitted)

Page 17 of17

Cytaty

Powiązane dokumenty

In order to study the existence of solutions we replace these two problems with their multivalued approximations and, for the first problem, we estabilish an existence result while

The convergence of difference schemes was proved first locally, next in the unbounded case for differential problems [2], and finally for differential-functional systems using a

Pinch [P] which, though based on the theory of linear forms in logarithms of algebraic num- bers, does not require high precision computations with algebraic numbers and

In this paper we study the periodic-Neumann boundary value problem for a class of nonlinear parabolic equations1. We prove a new uniqueness result and study the structure of the set

But there are some difficulties connected with the choice of a suitable version of this theorem, caused by the occurrence of singular points and by the necessity of integrating

These notions were intended as a refinement to the Nielsen theory of periodic orbits in that isotopy is much stronger than homotopy for closed curves in 3-manifolds, hence an

We first notice that if the condition (1.7) is satisfied then the a priori estimates for u − ε 1 (x) given in Corollary 3.3 can be modified so as to be independent of ε... Below

An infinite family of T -factorizations of complete graphs K 2n , where 2n = 56k and k is a positive integer, in which the set of vertices of T can be split into two subsets of the