Skip to main content

Advertisement

Solvability for second-order nonlocal boundary value problems with dim(kerM)=2

Article metrics

  • 1865 Accesses

Abstract

The existence of at least one solution to the second-order nonlocal boundary value problems on the real line is investigated by using an extension of Mawhin’s continuation theorem.

MSC: 34B10, 34B40, 34B15.

1 Introduction

Boundary value problems on an infinite interval arise quite naturally in the study of radially symmetric solutions of nonlinear elliptic equations and in various applications such as an unsteady flow of gas through a semi-infinite porous medium, theory of drain flows and plasma physics. For an extensive collection of results to boundary value problems on unbounded domains, we refer the reader to a monograph by Agarwal and O’Regan [1]. The study of nonlocal elliptic boundary value problems was investigated by Bicadze and Samarskiĭ [2], and later continued by Il’in and Moiseev [3] and Gupta [4]. Since then, the existence of solutions for nonlocal boundary value problems has received a great deal of attention in the literature. For more recent results, we refer the reader to [5]–[22] and the references therein.

In this paper, we consider the following second-order nonlinear differential equation with integral boundary conditions:

{ ( c φ p ( u ) ) ( t ) = f ( t , u ( t ) , u ( t ) ) , a.e.  t ( , ) , lim t ( c φ p ( u ) ) ( t ) = g ( s ) ( c φ p ( u ) ) ( s ) d s , lim t ( c φ p ( u ) ) ( t ) = h ( s ) ( c φ p ( u ) ) ( s ) d s ,
(1)

where φ p (s):= | s | p 2 s, p>1, f: ( , ) 3 (,) is a Carathéodory function, i.e., f=f(t,u,v) is Lebesgue measurable in t for all (u,v) ( , ) 2 and continuous in (u,v) for almost all t(,). Throughout this paper, we assume that the following assumptions hold:

(H1)g,h L 1 (,) satisfy g(s)ds= h(s)ds=1;

(H2)c:(,)(0,) is a continuous function which satisfy φ p 1 ( 1 c ) L loc 1 (,) L 1 (,);

(H3) let w(t):= 0 t φ p 1 ( 1 c ( s ) )ds, and there exist nonnegative measurable functions α, β and γ such that ( 1 + | w | ) p 1 α,β/c,γ L 1 (,) and

| f ( t , u , v ) | α(t) | u | p 1 +β(t) | v | p 1 +γ(t),a.e. t(,);

(H4) there exists a function k(t) such that (1+|w()|) e k ( ) L 1 (,) and

:= a 11 a 22 a 12 a 21 0,

where a 11 := Q 2 (w() e k ( ) ), a 12 := Q 1 (w() e k ( ) ), a 21 := Q 2 ( e k ( ) ), a 22 := Q 1 ( e k ( ) ), and Q 1 , Q 2 : L 1 (,)(,) will be defined in Section 3.

A boundary value problem is called a resonance one if the corresponding homogeneous boundary value problem has a non-trivial solution. Resonance problems can be expressed as an abstract equation Lx=Nx, where L is a noninvertible operator. When L is linear, Mawhin’s continuation theorem [23] is an efficient tool in finding solutions for these problems. However, it is not suitable for the case L is nonlinear. Recently, Ge and Ren [24] extended Mawhin’s continuation theorem from the case of linear L to the case of quasi-linear L. The purpose of this paper is to establish the sufficient conditions for the existence of solutions to the problem (1) on the real line at resonance with dim(kerL)=2 by using an extension of Mawhin’s continuation theorem [24].

2 Preliminaries

In this section, we recall some definitions and theorems. Let X and Y be two Banach spaces with the norms X and Y , respectively.

Definition 2.1

A continuous operator M:XdomMY is said to be quasi-linear if

  1. (i)

    ImM:=M(XdomM) is a closed subset of Y;

  2. (ii)

    KerM:={xXdomM:Mx=0} is linearly homeomorphic to ( , ) n for some n<.

Definition 2.2

Let M:XdomMY be a quasi-linear operator. Let X 1 =KerM and ΩX be an open and bounded set with the origin θ X Ω. Then N λ : Ω ¯ Y, λ[0,1] is said to be M-compact in Ω ¯ if N λ : Ω ¯ Y, λ[0,1] is a continuous operator, and there exist a vector subspace Y 1 of Y satisfying dim Y 1 =dim X 1 and an operator R: Ω ¯ ×[0,1] X 2 being continuous and compact such that, for λ[0,1],

  1. (i)

    (IQ) N λ ( Ω ¯ )ImM(IQ)Y;

  2. (ii)

    Q N λ x= θ Y , λ(0,1) Q N 1 x= θ Y ;

  3. (iii)

    R(,0) is the zero operator and R(,λ) | Σ λ =(IP) | Σ λ , where Σ λ ={x Ω ¯ :Mx= N λ x};

  4. (iv)

    M[P+R(,λ)]=(IQ) N λ .

Here, X 2 is a complement space of X 1 in X, θ Y is the origin of Y and P:X X 1 , Q:Y Y 1 are projections.

Now, we give an extension of Mawhin’s continuation theorem [24].

Theorem 2.3

LetΩXbe an open and bounded set with θ X Ω. Suppose thatM:XdomMYis a quasi-linear operator and N λ : Ω ¯ Y, λ[0,1]is M-compact. In addition, if the following conditions hold:

(A1)Mx N λ xfor every(u,λ)(domMΩ)×(0,1);

(A2)deg{JQ N 1 ,ΩKerM,0}0, whereJ:ImQKerMis a homeomorphism withJ( θ X )= θ Y ,

then the abstract equationMx= N 1 xhas at least one solution in Ω ¯ .

Finally, we give a theorem which is useful to show the compactness of operators defined on an infinite interval.

Theorem 2.4

[1]

Let Z be the space of all bounded continuous functions on(,)andSZ. Then S is relatively compact in Z if the following conditions hold:

  1. (i)

    S is bounded in Z;

  2. (ii)

    S is equicontinuous on any compact interval of (,);

  3. (iii)

    S is equiconvergent at ±∞, that is, given ϵ>0, there exists a constant T=T(ϵ)>0 such that |ϕ(t)ϕ()|<ϵ (respectively, |ϕ(t)ϕ()|<ϵ) for all t>T (respectively, t<T) and all ϕS.

3 Main result

Let X be the set of the functions u C 1 (,) such that

u 1 + | w | , φ p 1 (c) u L (,),

where w is the function in the assumption (H3). Then X is a Banach space equipped with a norm u X = u 1 + u 2 , where

u 1 = sup t ( , ) | u ( t ) | 1 + | w ( t ) | and u 2 = sup t ( , ) | ( φ p 1 ( c ) u ) ( t ) | .

Let Y denote the Banach space L 1 (,) equipped with a usual norm

h Y = | h ( s ) | ds.

Remark 3.1

  1. (1)

    It is well known that, for any u,v(,) and q>0,

    | u + v | q max { 1 , 2 q 1 } ( | u | q + | v | q ) .

Thus, φ p 1 (u+v) α p ( φ p 1 (u)+ φ p 1 (v)) for all u,v0, where α p :=max{1, 2 2 p p 1 }.

  1. (2)

    Since φ p 1 ( 1 c ) L loc 1 (,)Y, then w is a continuous function which satisfies lim t w(t)= and lim t w(t)=.

  2. (3)

    For any continuous functions w(t), we can choose a function k(t) which satisfies (1+|w()|) e k ( ) Y. For example, put k(t)= 0 t (1+|w(s)|)ds, then (1+|w()|) e k ( ) Y.

Define M:XdomMY by Mu= ( c φ p ( u ) ) , where

dom M = { u : ( c φ p ( u ) ) Y , lim t ( c φ p ( u ) ) ( t ) = g ( s ) ( c φ p ( u ) ) ( s ) d s , and lim t ( c φ p ( u ) ) ( t ) = h ( s ) ( c φ p ( u ) ) ( s ) d s } .

Then M:XdomMY is continuous. Let Ω be an open bounded subset of X such that domM Ω ¯ . For λ[0,1], define N λ :ΩY by N λ x=λf(,x, x ). By (H3) and the Lebesgue dominated convergence theorem, N λ is continuous. Denote N 1 by N. Then problem (1) is equivalent to Mx=Nx, xdomM. Define Q 1 , Q 2 :Y(,) by

Q 1 (y):= g(s) s y(τ)dτds, Q 2 (y):= h(s) s y(τ)dτds.

Then Q 1 , Q 2 :Y(,) are continuous.

Lemma 3.2

Assume that (H1) and (H2) hold. Then the operatorM:XdomMYis quasi-linear. Moreover, KerM={a+bw:a,b(,)}andImM={yY: Q 1 (y)= Q 2 (y)=0}.

Proof

Clearly, KerM={a+bw:a,b(,)}, and it is linearly homeomorphic to ( , ) 2 . Next, we show that

ImM= { y Y : Q 1 ( y ) = Q 2 ( y ) = 0 } .

Let yImM. Then there exists xXdomM such that

( c φ p ( x ) ) (t)=y(t),t(,).

For t(,),

( c φ p ( x ) ) (t)= ( c φ p ( x ) ) ()+ t y(s)ds

and

g(s) ( c φ p ( x ) ) (s)ds= ( c φ p ( x ) ) ()+ g(s) s y(τ)dτds.

Thus Q 1 (y)=0. In a similar manner, Q 2 (y)=0.

On the other hand, let yY satisfying Q 1 (y)= Q 2 (y)=0. Take

x(t)= 0 t φ p 1 ( 1 c ( s ) ) φ p 1 ( 0 s y ( τ ) d τ ) ds.

Then xXdomM, and ( c φ p ( x ) ) =yImM. Thus, ImM={yY: Q 1 (y)= Q 2 (y)=0}. Since Q 1 , Q 2 :Y(,) are continuous, ImM is closed in Y. Consequently, M is a quasi-linear operator.

Let T 1 , T 2 :YY be linear operators which are defined as follows:

T 1 y= 1 ( a 11 Q 1 ( y ) + a 12 Q 2 ( y ) ) e k ( )

and

T 2 y= 1 ( a 21 Q 1 ( y ) + a 22 Q 2 ( y ) ) e k ( ) ,

where a i j (i,j=1,2) are the constants in the assumption in (H4). Then, by direct calculations,

T 1 ( T 1 y)= T 1 y, T 1 ( ( T 2 y ) w ) =0, T 2 ( T 1 y)=0and T 2 ( ( T 2 y ) w ) = T 2 y.

Define the bounded linear operators Q:Y Y 1 and P:X X 1 by

Q(y)= T 1 y+( T 2 y)w,P(x)=x(0)+ ( φ p 1 ( c ) x ) (0)w,

where X 1 :=KerM and Y 1 :=ImQ={(a+bw()) e k ( ) :a,b(,)}. Then Q:Y Y 1 , P:X X 1 are projections, and dim Y 1 =dim X 1 =2. By (H4), 0, and it follows from Lemma 3.2 that ImM=KerQ. □

Lemma 3.3

Assume that (H1)-(H4) hold. Assume that Ω is an open bounded subset of X such thatdomM Ω ¯ . Then N λ : Ω ¯ Y, λ[0,1]is M-compact on Ω ¯ .

Proof

Let X 2 :=KerP. Then X 2 is a complement space of X 1 in X, i.e., X= X 1 X 2 . Define R: Ω ¯ ×[0,1] X 2 , for t(,), by

R ( x , λ ) ( t ) = 0 t φ p 1 ( 1 c ( s ) ) [ φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 s ( I Q ) N x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) ] d s .

Since Ω is bounded, there exists a constant r>0 such that x X r for any x Ω ¯ . For x Ω ¯ , and for almost all t(,), by (H3)

| ( N x ) ( t ) | = | f ( t , x ( t ) , x ( t ) ) | ( 1 + | w ( t ) | ) p 1 α ( t ) ( | x ( t ) | 1 + w ( t ) ) p 1 + β c ( t ) | ( φ p 1 ( c ) x ) ( t ) | p 1 + γ ( t ) ( ( 1 + | w ( t ) | ) p 1 α ( t ) + β c ( t ) ) x X p 1 + γ ( t ) ,
(2)

which implies that

N x Y r p 1 ( 1 + | w | ) p 1 α + β c Y + γ Y =: l r .

Since | Q 1 (Nx)| N x Y and | Q 2 (Nx)| N x Y , for x Ω ¯ ,

| Q ( N x ) ( t ) | | T 1 ( N x ) ( t ) | + | ( T 2 ( N x ) ( t ) ) w ( t ) | 1 | | ( | a 11 | | Q 1 ( N x ) | + | a 12 | | Q 2 ( N x ) | + ( | a 21 | | Q 1 ( N x ) | + | a 22 | | Q 2 ( N x ) | ) | w ( t ) | ) e k ( t ) N x Y ( M 1 + M 2 | w ( t ) | ) e k ( t ) l r ( M 1 + M 2 | w ( t ) | ) e k ( t ) .
(3)

Here,

M 1 := 1 | | [ | a 11 | + | a 12 | ] and M 2 := 1 | | [ | a 21 | + | a 22 | ] .

Thus

Q ( N x ) Y D N x Y ,
(4)

where

D:= ( M 1 + M 2 | w ( τ ) | ) e k ( τ ) dτ.
(5)

First, we prove that R: Ω ¯ ×[0,1] X 2 is compact by using Theorem 2.4. Let Z=C(,) L (,) with the usual sup norm. For x Ω ¯ ,

sup t ( , ) | R ( x , λ ) ( t ) | 1 + | w ( t ) | ( φ p 1 ( | ( c φ p ( x ) ) ( 0 ) | + | N x ( τ ) | + | Q ( N x ) ( τ ) | d τ ) + | ( φ p 1 ( c ) x ) ( 0 ) | ) ( α p + 1 ) r + α p ( ( 1 + D ) N x Y ) 1 p 1 ( α p + 1 ) r + α p ( ( 1 + D ) l r ) 1 p 1

and

sup t ( , ) | ( φ p 1 ( c ) R ( x , λ ) ) ( t ) | = sup t ( , ) | φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 t ( I Q ) N x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) | ( α p + 1 ) r + α p ( ( 1 + D ) N x Y ) 1 p 1 ( α p + 1 ) r + α p ( ( 1 + D ) l r ) 1 p 1 .

Here α p is the constant in Remark 3.1(1). Thus { R ( x , λ ) 1 + | w ( x ) | :x Ω ¯ } and { φ p 1 (c)R ( x , λ ) :x Ω ¯ } are bounded in Z.

Let T>0 and let ϵ>0 be given. First, for any t 1 , t 2 [0,T] with t 1 < t 2 , we have

| R ( x , λ ) ( t 1 ) 1 + w ( t 1 ) R ( x , λ ) ( t 2 ) 1 + w ( t 2 ) | w ( t 2 ) w ( t 1 ) ( 1 + w ( t 1 ) ) ( 1 + w ( t 2 ) ) | w ( t 1 ) | ( ( α p + 1 ) r + α p ( ( 1 + D ) N x Y ) 1 p 1 ) + 1 1 + w ( t 2 ) ( w ( t 2 ) w ( t 1 ) ) ( ( α p + 1 ) r + α p ( ( 1 + D ) N x Y ) 1 p 1 ) 2 ( w ( t 2 ) w ( t 1 ) ) ( ( α p + 1 ) r + α p ( ( 1 + D ) N x Y ) 1 p 1 )

and

| ( φ p 1 ( c ) R ( x , λ ) ) ( t 1 ) ( φ p 1 ( c ) R ( x , λ ) ) ( t 2 ) | = | φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 t 1 ( I Q ) N x ( τ ) d τ ) φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 t 2 ( I Q ) N x ( τ ) d τ ) | .

By (2) and (3), there exists zY such that

| ( I Q ) N x | zfor all x Ω ¯ ,
(6)

and since φ p 1 and w are uniformly continuous on a compact interval in (,), there exists δ 1 >0 such that if | t 1 t 2 |< δ 1 with t 1 , t 2 [0,T], then

| R ( x , λ ) ( t 1 ) 1 + | w ( t 1 ) | R ( x , λ ) ( t 2 ) 1 + | w ( t 2 ) | |< ϵ 2

and

| ( φ p 1 ( c ) R ( x , λ ) ) ( t 1 ) ( φ p 1 ( c ) R ( x , λ ) ) ( t 2 ) | < ϵ 2 .

In a similar manner, there exists δ 2 >0 such that if | t 1 t 2 |< δ 2 with t 1 , t 2 [T,0], then

| R ( x , λ ) ( t 1 ) 1 + | w ( t 1 ) | R ( x , λ ) ( t 2 ) 1 + | w ( t 2 ) | |< ϵ 2

and

| ( φ p 1 ( c ) R ( x , λ ) ) ( t 1 ) ( φ p 1 ( c ) R ( x , λ ) ) ( t 2 ) | < ϵ 2 .

Letting δ=min{ δ 1 , δ 2 }>0, if | t 1 t 2 |<δ with t 1 , t 2 [T,T], then

| R ( x , λ ) ( t 1 ) 1 + | w ( t 1 ) | R ( x , λ ) ( t 2 ) 1 + | w ( t 2 ) | |<ϵ

and

| ( φ p 1 ( c ) R ( x , λ ) ) ( t 1 ) ( φ p 1 ( c ) R ( x , λ ) ) ( t 2 ) | <ϵ.

Consequently, { R ( x , λ ) 1 + | w ( x ) | :x Ω ¯ } and { φ p 1 (c)R ( x , λ ) :x Ω ¯ } are equicontinuous on any compact intervals in (,).

For x Ω ¯ , by L’Hôspital’s rule,

lim t R ( x , λ ) ( t ) 1 + | w ( t ) | = lim t 1 1 + | w ( t ) | 0 t φ p 1 ( 1 c ( s ) ) [ φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 s ( I Q ) N x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) ] d s = lim t φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 t ( I Q ) N x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) = φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 ( I Q ) N x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 )

and

lim t ( φ p 1 ( c ) R ( x , λ ) ) ( t ) = φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 ( I Q ) N x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) .

In a similar manner,

lim t R ( x , λ ) ( t ) 1 + | w ( t ) | = φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 ( I Q ) N x ( τ ) d τ ) + ( φ p 1 ( c ) x ) ( 0 )

and

lim t ( φ p 1 ( c ) R ( x , λ ) ) ( t ) = φ p 1 ( ( c φ p ( x ) ) ( 0 ) λ 0 ( I Q ) N x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) .

By (6), we conclude that { R ( x , λ ) 1 + | w ( x ) | :xΩ} and { φ p 1 (c)R ( x , λ ) :xΩ} are equiconvergent at ±∞. Thus, R: Ω ¯ ×[0,1] X 2 is compact in view of Theorem 2.4.

Next, we prove that R: Ω ¯ ×[0,1] X 2 is continuous. Let {( x n , λ n )} be a sequence in Ω ¯ ×[0,1] such that x n x in X and λ n λ in (,) as n. Then { x n } is bounded in X and x n (t)x(t) pointwise as n. Since R is compact, there exists a subsequence {( x n k , λ n k )} of {( x n , λ n )} such that R( x n k , λ n k )(t)L in X as n k . By the Lebesgue dominated convergence theorem, R( x n , λ n )(t)R(x,λ)(t) as n. Thus, LR(x,λ). By a standard argument, R: Ω ¯ ×[0,1] X 2 is continuous.

Finally, we show that (i)-(iv) hold in Definition 2.2. Since Q(IQ) N λ ( Ω ¯ )=0, (IQ) N λ ( Ω ¯ )KerQ=ImM. For yImM, Qy=0, and y=(IQ)y(IQ)Y. Consequently, (IQ) N λ ( Ω ¯ )ImM(IQ)Y. Since N λ x=λNx for any x Ω ¯ ,

Q N λ x=0,λ(0,1)QNx=0,

and R(,0)= θ X . For x Σ λ ={x Ω ¯ :Mx= N λ x}, N λ x= ( c φ p ( x ) ) ImM=KerQ. Then, for xX and t(,),

R ( x , λ ) ( t ) = 0 t φ p 1 ( 1 c ( s ) ) [ φ p 1 ( ( c φ p ( x ) ) ( 0 ) + 0 s ( I Q ) N λ x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) ] d s = 0 t φ p 1 ( 1 c ( s ) ) [ φ p 1 ( ( c φ p ( x ) ) ( 0 ) + 0 s N λ x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) ] d s = 0 t φ p 1 ( 1 c ( s ) ) [ ( φ p 1 ( c ) x ) ( s ) ( φ p 1 ( c ) x ) ( 0 ) ] d s = x ( t ) ( x ( 0 ) + ( φ p 1 ( c ) x ) ( 0 ) w ( t ) ) = ( I P ) x ( t ) .

On the other hand, for xX and t(,),

M [ P x + R ( x , λ ) ] ( t ) = M [ x ( 0 ) + ( φ p 1 ( c ) x ) ( 0 ) w ( t ) + 0 t φ p 1 ( 1 c ( s ) ) [ φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 s ( I Q ) N x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) ] d s ] = ( I Q ) N λ x ( t ) .

Thus, N λ is M-compact on Ω ¯ . □

Now, we give the main result in this paper.

Theorem 3.4

Assume that (H1)-(H4) hold. Assume also that the following hold:

(H5)there exist positive constants A and B such that if|x(t)|>Afor everyt[B,B]or|(c φ p ( x ))(t)|>Afor everyt(,), thenQ(Nx)0, i.e., either Q 1 (Nx)0or Q 2 (Nx)0;

(H6)there exists a positive constant C such that if|a|>Cor|b|>C, then either

  1. (1)

    a Q 1 (N(a+bw()))+b Q 2 (N(a+bw()))<0 or

  2. (2)

    a Q 1 (N(a+bw()))+b Q 2 (N(a+bw()))>0.

Then problem (1) has at least one solution in X provided that

( 1 + | w | ) p 1 α Y + β c Y < ( 1 2 α p 3 + ( 4 + | w ( B ) | + 2 ( 1 + D ) 1 p 1 ) α p 2 ) p 1 .
(7)

Here, D is the constant defined in (5).

Proof

We divide the proof into three steps.

Step 1. Let

Ω 1 = { x dom M : M x = N λ x ,  for some  λ ( 0 , 1 ) } .

We will prove that Ω 1 is bounded. For x Ω 1 , Mx= N λ xImM=KerQ.

Thus

Q 1 (Nx)= Q 2 (Nx)=0.

By (H5), there exist t 0 [B,B] and t 1 (,) such that

| x ( t 0 ) | Aand | ( c φ p ( x ) ) ( t 1 ) | A,

which imply that

| ( c φ p ( x ) ) ( 0 ) | = | ( c φ p ( x ) ) ( t 1 ) ( c φ p ( x ) ) ( t 1 ) + ( c φ p ( x ) ) ( 0 ) | | ( c φ p ( x ) ) ( t 1 ) | + | 0 t 1 ( c φ p ( x ) ) ( s ) d s | A + M x Y A + N x Y ,

and |( φ p 1 (c) x )(0)| ( A + N x Y ) 1 p 1 α p A 1 p 1 + α p N x Y 1 p 1 . Then we have

| x ( 0 ) | = | x ( t 0 ) 0 t 0 φ p 1 ( 1 c ( s ) ) φ p 1 [ ( c φ p ( x ) ) ( t 1 ) + t 1 s ( c φ p ( x ) ) ( τ ) d τ ] d s | | x ( t 0 ) | + | w ( B ) | [ φ p 1 ( | ( c φ p ( x ) ) ( t 1 ) | + 0 | ( c φ p ( x ) ) ( τ ) | d τ ) ] A + | w ( B ) | ( α p A 1 p 1 + α p N x Y 1 p 1 ) .

Thus,

P x X = P x 1 + P x 2 | x ( 0 ) | + 2 | ( φ p 1 ( c ) x ) ( 0 ) | A + α p ( 2 + | w ( B ) | ) A 1 p 1 + α p ( 2 + | w ( B ) | ) N x Y 1 p 1 .

On the other hand, by (4),

| R ( x , λ ) ( t ) | 1 + | w ( t ) | = 1 1 + | w ( t ) | | 0 t φ p 1 ( 1 c ( s ) ) [ φ p 1 ( ( c φ p ( x ) ) ( 0 ) + λ 0 s ( I Q ) N x ( τ ) d τ ) ( φ p 1 ( c ) x ) ( 0 ) ] d s | ( α p + 1 ) φ p 1 ( | ( c φ p ( x ) ) ( 0 ) | ) + α p φ p 1 ( N x Y + Q N x Y ) ( α p + 1 ) ( A + N x Y ) 1 p 1 + α p ( 1 + D ) 1 p 1 N x Y 1 p 1 = α p ( α p + 1 ) A 1 p 1 + α p ( α p + 1 + ( 1 + D ) 1 p 1 ) N x Y 1 p 1

and

| ( φ p 1 ( c ) ( R ( x , λ ) ) ) ( t ) | α p ( α p + 1 ) A 1 p 1 + α p ( α p + 1 + ( 1 + D ) 1 p 1 ) N x Y 1 p 1 .

Thus,

R ( x , λ ) X 2 α p ( α p +1) A 1 p 1 +2 α p ( α p + 1 + ( 1 + D ) 1 p 1 ) N x Y 1 p 1 .

It follows that

x X = P x + ( I P ) x X P x X + ( I P ) x X = P x X + R ( x , λ ) X A + α p ( 2 + | w ( B ) | ) A 1 p 1 + α p ( 2 + | w ( B ) | ) N x Y 1 p 1 + 2 α p ( α p + 1 ) A 1 p 1 + 2 α p ( α p + 1 + ( 1 + D ) 1 p 1 ) N x Y 1 p 1 A + α p ( 2 α p + 4 + | w ( B ) | ) A 1 p 1 + α p 2 ( 2 α p + 4 + | w ( B ) | + 2 ( 1 + D ) 1 p 1 ) γ Y 1 p 1 + α p 2 ( 2 α p + 4 + | w ( B ) | + 2 ( 1 + D ) 1 p 1 ) ( ( 1 + | w | ) p 1 α Y + β c Y ) 1 p 1 x X .

By (7), Ω 1 is bounded.

Step 2. Define a homeomorphism J:ImQKerM by

J ( ( a + b w ( ) ) e k ( ) ) = a 22 a a 12 b+( a 21 a+ a 11 b)w().

Assume (H6)(1) holds, i.e., there exists a positive constant C such that if |a|>C or |b|>C, then a Q 1 (N(a+bw()))+b Q 2 (N(a+bw()))<0. Let

Ω 2 = { x ker M : λ x + ( 1 λ ) J Q N x = 0 ,  for some  λ [ 0 , 1 ] } .

Let x Ω 2 . Then x=a+bw for some a,b(,). If λ=0, JQN(a+bw())=0. Since J is homeomorphism, QN(a+bw())=0. By (H6), we obtain |a|C and |b|C. If λ=1, then a=b=0.

For λ(0,1), by λx=(1λ)JQNx, we obtain

λa=(1λ) Q 1 ( N ( a + b w ( ) ) ) ,λb=(1λ) Q 2 ( N ( a + b w ( ) ) ) .

If |a|>C or |b|>C, then, by (H6)(1), we obtain

λ ( a 2 + b 2 ) =(1λ) ( a Q 1 N ( a + b w ( ) ) + b Q 2 N ( a + b w ( ) ) ) <0,

which is a contradiction. Thus, Ω 2 is bounded.

In the case that (H6)(2) holds, we take

Ω 2 = { x ker M : λ x + ( 1 λ ) J Q N x = 0  for some  λ [ 0 , 1 ] } ,

and it follows that Ω 2 is bounded in a similar manner.

Step 3. Take an open bounded set Ω Ω ¯ 1 Ω ¯ 2 {0} in X. By Step 1,

(A1)Mu N λ u for every (u,λ)(domMΩ)×(0,1).

Now we will show that

(A2)deg(JQN,ΩkerM,0)0.

Let H(x,λ)=±λx+(1λ)JQNx. By Step 2, we know that H(x,λ)0, for every (x,λ)(kerMΩ)×[0,1]. Thus, by the homotopy property of the degree, we obtain

deg ( J Q N , Ω ker M , 0 ) = deg ( H ( , 0 ) , Ω ker M , 0 ) = deg ( H ( , 1 ) , Ω ker M , 0 ) = deg ( ± I , Ω ker M , 0 ) = ± 1 0 .

By Theorem 2.3, Mx=Nx has at least one solution in domMΩ, and consequently problem (1) has at least one solution in X. □

4 Example

Consider the following second-order nonlinear differential equation:

{ ( | u | u ) ( t ) = f ( t , u ( t ) , u ( t ) ) , a.e.  t ( , ) , lim t ( | u | u ) ( t ) = g ( s ) ( | u | u ) ( s ) d s , lim t ( | u | u ) ( t ) = h ( s ) ( | u | u ) ( s ) d s ,
(8)

where

g(t)={ 2 t , t [ 1 , 0 ] , 0 , otherwise , h(t)={ 2 t , t [ 0 , 1 ] , 0 , otherwise .

Define f: ( , ) 3 (,) by f(t,u,v)=α(t)u+β(t)v+ γ ¯ (t), where

α ( t ) = { 10 3 e t , t [ 1 , 0 ] , 0 , otherwise , β ( t ) = { 10 4 e t , t 1 , 0 , otherwise , γ ¯ ( t ) = { e t + 15 e 1 6 , t [ 1 , 0 ] , 0 , otherwise .

Then

| f ( t , u , v ) | α ( t ) | u | + β ( t ) | v | + | γ ¯ ( t ) | α ( t ) ( | u | 2 + 1 ) + β ( t ) ( | v | 2 + 1 ) + | γ ¯ ( t ) | = α ( t ) | u | 2 + β ( t ) | v | 2 + γ ( t ) ,

where γ(t)=α(t)+β(t)+| γ ¯ (t)|. Since c(t)=1 for t(,) and p=3, w(t)=t, and thus (H1), (H2), and (H3) hold.

For yY,

Q 1 (y)= 1 y(τ)dτ+ 1 0 y(τ) τ 2 dτ

and

Q 2 (y)= 1 y(τ)dτ+ 0 1 y(τ) τ 2 dτ.

Take k(t)=|t|, then

a 11 =14 e 1 +6, a 12 =14 e 1 +6, a 21 =4 e 1 2, a 22 =4 e 1 +2,

and

= a 11 a 22 a 12 a 21 = ( 14 e 1 + 6 ) ( 8 e 1 + 4 ) >0.

Thus, (H4) holds.

Take B=1 and A=1. If |x(t)|>1, for t[1,1], then | Q 1 (Nx)|=| 10 3 1 0 t 2 e | t | x(t)dt|>| 10 3 1 0 t 2 e | t | dt|= 10 3 (5 e 1 +2)>0. If |(| x | x )(t)|>1 for t(,), then | x (t)|>1 for t(,), and

| Q 2 ( N x ) | =| 10 4 1 e | t | x (t)dt|>| 10 4 1 e | t | dt|= 10 4 e 1 >0.

Thus, (H5) holds.

For any C>0, if |a|>C or |b|>C, then

a Q 1 ( N ( a + b t ) ) + b Q 2 ( N ( a + b t ) ) = 10 3 ( 5 e 1 + 2 ) a 2 + 10 3 ( 16 e 1 6 ) a b + 10 4 e 1 b 2 = 10 3 ( 5 e 1 + 2 ) ( a + 8 e 1 3 5 e 1 + 2 b ) 2 + 10 4 ( e 1 10 ( 8 e 1 3 ) 2 5 e 1 + 2 ) b 2 > 0 .

Thus, (H6)(2) is satisfied.

Since p=3, α p =1, B=1, and D= 1 2 e 1 + 1 + 1 7 e 1 + 3 , we have

( 1 + | w | ) 2 α Y + β c Y < 4 × 10 4 < ( 1 2 + ( 5 + 2 ( 1 + 1 2 e 1 + 1 + 1 7 e 1 + 3 ) 1 2 ) ) 2

and (7) holds. Consequently, there exists at least one solution to problem (8) in view of Theorem 3.4.

References

  1. 1.

    Agarwal RP, O’Regan D: Infinite Interval Problems for Differential, Difference and Integral Equations. Kluwer Academic, Dordrecht; 2001.

  2. 2.

    Bicadze AV, Samarskiĭ AA: Some elementary generalizations of linear elliptic boundary value problems. Dokl. Akad. Nauk SSSR 1969, 185: 739-740.

  3. 3.

    Il’in VA, Moiseev EI: A nonlocal boundary value problem of the second kind for the Sturm-Liouville operator. Differ. Equ. 1987, 23: 979-987.

  4. 4.

    Gupta CP: Solvability of a three-point nonlinear boundary value problem for a second order ordinary differential equation. J. Math. Anal. Appl. 1992, 168(2):540-551. 10.1016/0022-247X(92)90179-H

  5. 5.

    Chen Y, Lv Z: Solvability of fractional-order multi-point boundary-value problems at resonance on the half-line. Electron. J. Differ. Equ. 2012., 2012: 10.1186/1687-1847-2012-230

  6. 6.

    Djebali S, Mebarki K: System of singular second-order differential equations with integral condition on the positive half-line. Electron. J. Qual. Theory Differ. Equ. 2013., 2013: 10.1186/1687-1847-2013-50

  7. 7.

    Du Z, Zhao B, Bai Z: Solvability of a third-order multipoint boundary value problem at resonance. Abstr. Appl. Anal. 2014., 2014:

  8. 8.

    Jeong J, Kim C-G, Lee EK:Solvability for nonlocal boundary value problems on a half line with dim(kerL)=2. Bound. Value Probl. 2014., 2014: 10.1186/s13661-014-0167-6

  9. 9.

    Jiang W: Solvability for p -Laplacian boundary value problem at resonance on the half-line. Bound. Value Probl. 2013., 2013: 10.1186/1687-2770-2013-207

  10. 10.

    Jiang W, Wang B, Wang Z:Solvability of a second-order multi-point boundary-value problems at resonance on a half-line with dimkerL=2. Electron. J. Differ. Equ. 2011., 2011:

  11. 11.

    Kim C-G: Existence and iteration of positive solutions for multi-point boundary value problems on a half-line. Comput. Math. Appl. 2011, 61(7):1898-1905. 10.1016/j.camwa.2011.02.023

  12. 12.

    Kim C-G: Existence of positive solutions for multi-point boundary value problem with strong singularity. Acta Appl. Math. 2010, 112(1):79-90. 10.1007/s10440-009-9554-x

  13. 13.

    Kim C-G, Lee EK: Multiple positive solutions for singular multi-point boundary-value problems with a positive parameter. Electron. J. Differ. Equ. 2014., 2014: 10.1186/1687-1847-2014-38

  14. 14.

    Kosmatov N: A multi-point boundary value problem with two critical conditions. Nonlinear Anal. 2006, 65(3):622-633. 10.1016/j.na.2005.09.042

  15. 15.

    Kosmatov N: Multi-point boundary value problems on an unbounded domain at resonance. Nonlinear Anal. 2008, 68(8):2158-2171. 10.1016/j.na.2007.01.038

  16. 16.

    Lian H, Ge W: Solvability for second-order three-point boundary value problems on a half-line. Appl. Math. Lett. 2006, 19(10):1000-1006. 10.1016/j.aml.2005.10.018

  17. 17.

    Massar M: Existence and multiplicity of solutions for nonlocal elliptic problems. Electron. J. Differ. Equ. 2013., 2013: 10.1186/1687-1847-2013-75

  18. 18.

    Meng F, Du Z: Solvability of a second-order multi-point boundary value problem at resonance. Appl. Math. Comput. 2009, 208(1):23-30. 10.1016/j.amc.2008.11.026

  19. 19.

    O’Regan D, Yan B, Agarwal RP: Solutions in weighted spaces of singular boundary value problems on the half-line. J. Comput. Appl. Math. 2007, 205(2):751-763. 10.1016/j.cam.2006.02.055

  20. 20.

    Xu J, Yang Z: Positive solutions for singular Sturm-Liouville boundary value problems on the half line. Electron. J. Differ. Equ. 2010., 2010:

  21. 21.

    Yan B, O’Regan D, Agarwal RP: Positive solutions for second order singular boundary value problems with derivative dependence on infinite intervals. Acta Appl. Math. 2008, 103(1):19-57. 10.1007/s10440-008-9218-2

  22. 22.

    Yang A, Miao C, Ge W: Solvability for second-order nonlocal boundary value problems with a p -Laplacian at resonance on a half-line. Electron. J. Qual. Theory Differ. Equ. 2009., 2009:

  23. 23.

    Mawhin J: Topological Degree Methods in Nonlinear Boundary Value Problems. Am. Math. Soc., Providence; 1979.

  24. 24.

    Ge W, Ren J: An extension of Mawhin’s continuation theorem and its application to boundary value problems with a p -Laplacian. Nonlinear Anal. 2004, 58(3-4):477-488. 10.1016/j.na.2004.01.007

Download references

Acknowledgements

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2012R1A1A1011225).

Author information

Correspondence to Eun Kyoung Lee.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

All authors contributed equally to the manuscript and read and approved the final draft.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Keywords

  • resonance
  • nonlocal boundary condition
  • solvability
  • infinite interval