Open Access

Infinitely many homoclinic solutions for a class of second-order Hamiltonian systems

Advances in Difference Equations20142014:161

https://doi.org/10.1186/1687-1847-2014-161

Received: 5 February 2014

Accepted: 12 May 2014

Published: 2 June 2014

Abstract

In this paper, we deal with the existence of infinitely many homoclinic solutions for a class of second-order Hamiltonian systems. By using the dual fountain theorem, we give some new criteria to guarantee that the second-order Hamiltonian systems have infinitely many homoclinic solutions. Some recent results are generalised and significantly improved.

MSC:34B08, 34B15, 34B37, 58E30.

Keywords

Hamiltonian systemshomoclinic solutionsvariational methodscritical points

1 Introduction

Consider the following second-order Hamiltonian systems:
u ¨ ( t ) L ( t ) u ( t ) + W u ( t , u ( t ) ) = 0 , t R ,
(1.1)

where L C ( R , R N × N ) is a symmetric matrix valued function, u R N and W C 1 ( R × R N , R ) . As usual, we say that a solution u of system (1.1) is homoclinic to zero if u C 2 ( R , R N ) , u 0 , u ( t ) 0 and u ˙ ( t ) 0 as | t | .

Inspired by the excellent monographs [1, 2], the existence of periodic solutions and homoclinic solutions for second-order Hamiltonian systems have been intensively studied in many recent papers via variational methods; see [330] and references therein. Recently, some researchers have begun to study the existence of solutions for second-order Hamiltonian systems with impulses by using some critical points theorems of [31, 32]; see [33, 34].

Homoclinic solutions of dynamical systems are very important in applications for a lot of reasons. They may be ‘organizing centers’ for the dynamics in their neighborhood. From their existence one may, under proper conditions, deduce the bifurcation behavior of periodic orbits or the existence of chaos nearby. In the past 20 years, with the aid of the variational methods, the existence and multiplicity of homoclinic solutions for system (1.1) have been extensively investigated by many authors; see [3, 4, 69, 1130] and references therein. Most of them treated the superquadratic case [8, 15, 18, 19, 21, 24, 25, 2729] treated subquadratic case and [9, 21, 30] treated asymptotically quadratic case. Particularly, Yang and Zhang [25] considered a superquadratic case and obtained system (1.1) has infinitely many homoclinic solutions by using the following conditions.

(A1) W ( t , u ) / | u | 2 + , as | u | uniformly for all t R , and

(A2) W u ( t , u ( t ) ) = o ( | u | ) , as | u | 0 uniformly for all t R .

Recently, Wei and Wang [22] dealt with a case that W ( t , u ) = F ( t , u ) + G ( t , u ) , where F ( t , u ) is subquadratic and G ( t , u ) is superquadratic, by using the following condition, they obtained system (1.1) has infinitely many homoclinic solutions.

(A3) There exists ν < 1 such that
l ( t ) | t | ν 2 as  | t | ,

where l ( t ) = inf | ξ | = 1 ( L ( t ) ξ , ξ ) .

In [23], Yang et al. obtained the following theorems by using the variant fountain theorem.

Theorem 1.1 ([[23], Theorem 1.2])

Assume that the following conditions are satisfied:

(C1) L C ( R , R N × N ) is a symmetric and positive definite matrix for all t R and there is a continuous function β : R R such that β ( t ) > 0 for all t R and ( L ( t ) u , u ) β ( t ) | u | 2 and β ( t ) as | t | .

( C 2 ) c 1 ( t ) | u | γ ( W u ( t , u ) , u ) , | W u ( t , u ) | c 2 ( t ) | u | γ 1 + c 3 ( t ) | u | σ 1 where c 1 , c 2 , c 3 : R R + are positive continuous functions such that c 1 , c 2 L 2 2 γ ( R , R + ) , c 3 L 2 2 σ ( R , R + ) and 1 < γ < 2 , 1 < σ < 2 are constants, W ( t , 0 ) = 0 , W ( t , u ) = W ( t , u ) for all ( t , u ) R × R N .

Then system (1.1) possesses infinitely many homoclinic solutions.

Theorem 1.2 ([[23], Theorem 1.3])

Assume that (C1) hold. Moreover, we assume that the following condition is satisfied:

( C 3 ) W ( t , u ) = m ( t ) | u | γ + d | u | q where m : R R + is a positive continuous function such that m L 2 2 γ ( R , R + ) and 1 < γ < 2 , d 0 , q > 2 are constants.

Then system (1.1) possesses infinitely many homoclinic solutions.

Motivated by the above facts, in this paper, we will improve and generalize some results in the references that we have mentioned above.

Now, we state our main results.

Theorem 1.3 Assume that (C1) hold. Moreover, we assume that the following conditions are satisfied:

(C2) W ( t , u ) = F ( t , u ) + G ( t , u ) where F ( t , 0 ) = 0 , G ( t , 0 ) = 0 and F , G C 1 ( R × R N , R ) are even in u.

(C3) c 1 ( t ) | u | γ ( F u ( t , u ) , u ) , | F u ( t , u ) | c 2 ( t ) | u | γ 1 + c 3 ( t ) | u | σ 1 where c 1 , c 2 , c 3 : R R + are positive continuous functions such that c 1 , c 2 L 2 2 γ ( R , R + ) , c 3 L 2 2 σ ( R , R + ) and 1 < γ < 2 , 1 < σ < 2 are constants.

(C4) G ( t , u ) 0 for all ( t , u ) R × R N and there exists μ > 2 such that
| G u ( t , u ) | a ( t ) + c ( t ) | u | μ 1 ,

where a , c : R R + are positive continuous functions such that a L 2 ( R , R + ) and c L ( R , R + ) .

(C5) There exist ρ > 2 and 1 < δ < 2 such that
ρ G ( t , u ) ( G u ( t , u ) , u ) h ( t ) | u | δ , ( t , u ) R × R N ,

where h : R R + is a positive continuous function such that h L 2 2 δ ( R , R + ) .

Then system (1.1) possesses infinitely many homoclinic solutions.

Remark 1.1 It is clear that there are many functions satisfying (C1) but do not satisfying (A3); see [28].

Remark 1.2 Obviously, Theorem 1.3 generalizes Theorem 1.2 in [23], Theorem 1.1 in [29] and Theorem 1.2 in [18]. In fact, let G ( t , u ) 0 , then Theorem 1.3 coincides with Theorem 1.2 in [23], and contains Theorem 1.1 in [29] and Theorem 1.2 in [18]. Furthermore, there are many functions W satisfying our Theorem 1.3 and not satisfying Theorem 1.2 in [23], Theorem 1.1 in [29] and Theorem 1.2 in [18]. For example, the function
W ( t , u ) = F ( t , u ) + G ( t , u ) ,
(1.2)

where F ( t , u ) = ( 1 1 + t 2 ) 1 3 | u | 4 3 + ( 1 1 + | t | ) 1 3 | u | 5 3 and G ( t , u ) = 1 1 + t 6 | u | 4 .

Remark 1.3 It is easy to see that there are many functions W satisfying the conditions of Theorem 1.3 but not satisfying (A1), (A2), the condition ( C 3 ) in Theorem 1.2 or conditions ( R 2 ) and ( R 4 ) in Theorem 1.1 in [22], for example, the function (1.2) since ( 1 1 + t 2 ) 1 3 0 , ( 1 1 + | t | ) 1 3 0 and 1 1 + t 6 | u | 4 0 as | t | .

Theorem 1.4 Assume that (C1)-(C3) hold. Moreover, we assume that the following conditions are satisfied:

(C6) There exist λ > 2 and 0 h 1 < β 0 ( λ 2 ) 2 such that
λ G ( t , u ) ( G u ( t , u ) , u ) h 1 | u | 2 , ( t , u ) R × R N ,

where β 0 = min t R β ( t ) .

(C7) There exists 0 < θ < β 0 such that lim u 0 G u ( t , u ) | u | < θ uniformly for t R .

(C8) G ( t , u ) 0 for all ( t , u ) R × R N and there exist h 2 > 0 and p 1 > 2 such that
| G u ( t , u ) | h 2 ( 1 + | u | p 1 1 ) .

Then system (1.1) possesses infinitely many homoclinic solutions.

Remark 1.4 Obviously, all the conditions in Theorem 1.4 are more general than those in Theorem 1.2. Therefore, Theorem 1.4 is a complement of Theorem 1.2. On the other hand, there are many functions W satisfying our Theorem 1.4 and not satisfying Theorem 1.2. For example, the function
W ( t , u ) = F ( t , u ) + G ( t , u ) ,
(1.3)

where F ( t , u ) = ( 1 1 + t 2 ) 1 3 | u | 4 3 + ( 1 1 + | t | ) 1 3 | u | 5 3 and G ( t , u ) = | u | 4 + β 0 4 | u | 2 . Thus, F ( t , u ) is subquadratic and G ( t , u ) is superquadratic. To the best of our knowledge, with the exception of [22, 23], the study of this case has received considerably less attention. Furthermore, it is easy to see that the function (1.3) does not satisfy Theorem 1.1 in [22].

Remark 1.5 In Theorem 1.4, there are many functions W satisfying (A1) and not satisfying (A2), for example, the function (1.3).

The remainder of this paper is organized as follows. In Section 2, some preliminary results are presented. In Section 3, we give the proofs of Theorems 1.3 and 1.4.

2 Preliminaries

We will present some definitions and lemmas that will be used in the proofs of our results. Let
E = { u H 1 ( R , R N ) : R [ | u ˙ ( t ) | 2 + ( L ( t ) u ( t ) , u ( t ) ) ] d t < + }
equipped with the norm
u = ( R [ | u ˙ ( t ) | 2 + ( L ( t ) u ( t ) , u ( t ) ) ] d t ) 1 2 , u X ,
(2.1)
and the inner product
u , v = R [ ( u ˙ ( t ) , v ˙ ( t ) ) + ( L ( t ) u ( t ) , v ( t ) ) ] d t , u , v X .
(2.2)
Then E is a Hilbert space with this inner product. Denote by E its dual space with the associated operator norm E . Note that E is continuously embedded in L p ( R , R N ) for all p [ 2 , + ] . Therefore, there exists a constant δ p > 0 such that
u p δ p u , u E ,
(2.3)

where p denotes the usual norm on L p ( R , R N ) .

Lemma 2.1 (see [13])

Suppose that L satisfies (C1). Then the embedding of E in L 2 ( R , R N ) is compact.

Lemma 2.2 Suppose that (C1), (C2), (C3) and (C4) are satisfied. If u k u in E, then W u ( t , u k ) W u ( t , u ) in L 2 ( R , R N ) .

Proof Assume that u k u . In view of (C2), (C3) and (C4), we get
| W u ( t , u k ) W u ( t , u ) | c 2 ( t ) ( | u k | γ 1 + | u | γ 1 ) + c 3 ( t ) ( | u k | σ 1 + | u | σ 1 ) + | G u ( t , u k ) G u ( t , u ) | c 2 ( t ) ( | u k | γ 1 + | u | γ 1 ) + c 3 ( t ) ( | u k | σ 1 + | u | σ 1 ) + c ( t ) ( | u k | μ 1 + | u | μ 1 ) + 2 a ( t ) c 2 ( t ) ( | u k u | γ 1 + 2 | u | γ 1 ) + c 3 ( t ) ( | u k u | σ 1 + 2 | u | σ 1 ) + c ( t ) [ 2 μ 2 | u k u | μ 1 + ( 2 μ 2 + 1 ) | u | μ 1 ] + 2 a ( t ) c 2 ( t ) ( | u k u | γ 1 + 2 | u | γ 1 ) + c 3 ( t ) ( | u k u | σ 1 + 2 | u | σ 1 ) + 2 μ 1 c ( t ) ( | u k u | μ 1 + | u | μ 1 ) + 2 a ( t ) ,
(2.4)
which yields
| W u ( t , u k ) W u ( t , u ) | 2 7 c 2 2 ( t ) ( | u k u | 2 γ 2 + 4 | u | 2 γ 2 ) + 7 c 3 2 ( t ) ( | u k u | 2 σ 2 + 4 | u | 2 σ 2 ) + 28 a 2 ( t ) + 7 × 2 2 μ 2 c 2 ( t ) ( | u k u | 2 μ 2 + | u | 2 μ 2 ) .
(2.5)
By virtue of (2.3), u k u and the Banach-Steinhaus Theorem, one has
sup k N u k M 1 , u M 1 ,
(2.6)
where M 1 > 0 is a constant. In view of 2 μ 4 > 0 , (2.5) and (2.6), we have
| W u ( t , u k ) W u ( t , u ) | 2 7 c 2 2 ( t ) ( | u k u | 2 γ 2 + 4 | u | 2 γ 2 ) + 7 c 3 2 ( t ) ( | u k u | 2 σ 2 + 4 | u | 2 σ 2 ) + 28 a 2 ( t ) + 7 × 2 2 μ 2 c 2 ( t ) ( | u k u | 2 μ 2 + | u | 2 μ 2 ) 7 c 2 2 ( t ) ( | u k u | 2 γ 2 + 4 | u | 2 γ 2 ) + 7 c 3 2 ( t ) ( | u k u | 2 σ 2 + 4 | u | 2 σ 2 ) + 28 a 2 ( t ) + 7 × 2 2 μ 2 c 2 ( t ) [ ( 2 M 1 ) 2 μ 4 | u k u | 2 + | u | 2 μ 2 ] .
(2.7)
Since u k u , then u k u in L 2 ( R , R N ) , passing to a subsequence if necessary, we have
u k u 2 0 as  k .
So it can be assumed that
k = 1 u k u 2 < + ,
which implies that u k ( t ) u ( t ) for almost every t R and
k = 1 | u k ( t ) u ( t ) | = e ( t ) L 2 ( R , R ) .
Then we have
| W u ( t , u k ) W u ( t , u ) | 2 7 c 2 2 ( t ) ( | e ( t ) | 2 γ 2 + 4 | u | 2 γ 2 ) + 7 c 3 2 ( t ) ( | e ( t ) | 2 σ 2 + 4 | u | 2 σ 2 ) + 28 a 2 ( t ) + 7 × 2 2 μ 2 c 2 ( t ) [ ( 2 M 1 ) 2 μ 4 | e ( t ) | 2 + | u | 2 μ 2 ] .
(2.8)
By (2.3), (2.8), 2 μ 2 > 2 and the Hölder inequality, we obtain
R | W u ( t , u k ) W u ( t , u ) | 2 d t 7 R c 2 2 ( t ) ( | e ( t ) | 2 γ 2 + 4 | u | 2 γ 2 ) d t + 7 R c 3 2 ( t ) ( | e ( t ) | 2 σ 2 + 4 | u | 2 σ 2 ) d t + 28 R a 2 ( t ) d t + 7 × 2 2 μ 2 c 2 R [ ( 2 M 1 ) 2 μ 4 | e ( t ) | 2 + | u | 2 μ 2 ] d t 7 c 2 2 2 γ 2 ( e 2 2 γ 2 + 4 u 2 2 γ 2 ) + 28 a 2 2 + 7 c 3 2 2 σ 2 ( e 2 2 σ 2 + 4 u 2 2 σ 2 ) + 7 × 2 2 μ 2 c 2 [ ( 2 M 1 ) 2 μ 4 e 2 2 + u 2 μ 2 2 μ 2 ] 7 c 2 2 2 γ 2 ( e 2 2 γ 2 + 4 δ 2 2 γ 2 u 2 γ 2 ) + 28 a 2 2 + 7 c 3 2 2 σ 2 ( e 2 2 σ 2 + 4 δ 2 2 σ 2 u 2 σ 2 ) + 7 × 2 2 μ 2 c 2 [ ( 2 M 1 ) 2 μ 4 e 2 2 + δ 2 μ 2 2 μ 2 u 2 μ 2 ] .

By using the Lebesgue dominated convergence theorem, the lemma is proved. □

Remark 2.1 Suppose that the condition (C4) is replaced by conditions (C7) and (C8), then we can obtain the same conclusion.

Define the functional Ψ on E by
Ψ ( u ) = 1 2 R [ | u ˙ ( t ) | 2 + ( L ( t ) u ( t ) , u ( t ) ) ] d t R W ( t , u ( t ) ) d t = 1 2 u 2 φ ( u ) ,
(2.9)

where φ ( u ) = R W ( t , u ( t ) ) d t .

Lemma 2.3 Under conditions (C1)-(C4), we have
Ψ ( u ) v = R [ ( u ˙ ( t ) , v ˙ ( t ) ) + ( L ( t ) u ( t ) , v ( t ) ) ] d t R ( W u ( t , u ( t ) ) , v ( t ) ) d t = R [ ( u ˙ ( t ) , v ˙ ( t ) ) + ( L ( t ) u ( t ) , v ( t ) ) ] d t φ ( u ) v
(2.10)
for any u , v E , which yields
Ψ ( u ) u = u 2 R ( W u ( t , u ( t ) ) , u ( t ) ) d t .
(2.11)

Moreover, Ψ C 1 ( E , R ) , φ : E E is compact and any critical point of Ψ on E is a classical solution for system (1.1) satisfying u C 2 ( R , R N ) , u ( t ) 0 and u ˙ ( t ) 0 as | t | .

Proof We first show that Ψ : E R . It follows from (C2), (C3), (C4), (2.3), μ > 2 and the Hölder inequality that
0 R W ( t , u ( t ) ) d t R ( 1 γ c 2 ( t ) | u ( t ) | γ + 1 σ c 3 ( t ) | u ( t ) | σ ) d t + R ( a ( t ) | u ( t ) | + c ( t ) | u ( t ) | μ ) d t 1 γ c 2 2 2 γ u 2 γ + 1 σ c 3 2 2 σ u 2 σ + a 2 u 2 + c u μ μ 1 γ δ 2 γ c 2 2 2 γ u γ + 1 σ δ 2 σ c 3 2 2 σ u σ + δ 2 a 2 u + c δ μ μ u μ .
Next, we prove that Ψ C 1 ( E , R ) . It is sufficient to show that φ C 1 ( E , R ) . At first, we will see that
φ ( u ) v = R ( W u ( t , u ( t ) ) , v ( t ) ) d t
(2.12)
for any u , v E . For any given u E , let us define Φ ( u ) : E R as follows:
Φ ( u ) v = R ( W u ( t , u ( t ) ) , v ( t ) ) d t , v E .
It is easy to see that Φ ( u ) is linear. In the following we show that Φ ( u ) is bounded. In fact, for any u E , by (C2), (C3), (C4), (2.3) and the Hölder inequality, we have
| Φ ( u ) v | = | R ( W u ( t , u ( t ) ) , v ( t ) ) d t | R c 2 ( t ) | u ( t ) | γ 1 | v ( t ) | d t + R c 3 ( t ) | u ( t ) | σ 1 | v ( t ) | d t + R a ( t ) | v ( t ) | d t + R c ( t ) | u ( t ) | μ 1 | v ( t ) | d t ( R c 2 2 ( t ) | u ( t ) | 2 γ 2 d t ) 1 2 v 2 + ( R c 3 2 ( t ) | u ( t ) | 2 σ 2 d t ) 1 2 v 2 + a 2 v 2 + c ( R | u ( t ) | 2 μ 2 d t ) 1 2 v 2 ( c 2 2 2 γ u 2 γ 1 + c 3 2 2 σ u 2 σ 1 + a 2 + c u 2 μ 2 μ 1 ) v 2 δ 2 ( δ 2 γ 1 c 2 2 2 γ u γ 1 + δ 2 σ 1 c 3 2 2 σ u σ 1 + a 2 + c δ 2 μ 2 μ 1 u μ 1 ) v .
Moreover, for any u , v E , by the mean value theorem, we have
R [ W ( t , u ( t ) + v ( t ) ) W ( t , u ( t ) ) ] d t = R ( W u ( t , u ( t ) + ϑ ( t ) v ( t ) ) , v ( t ) ) d t ,
where ϑ ( t ) ( 0 , 1 ) . Thus, by Lemma 2.2 and the Hölder inequality, one has
R ( W u ( t , u ( t ) + ϑ ( t ) v ( t ) ) W u ( t , u ( t ) ) , v ( t ) ) d t 0 as  v 0 .
Suppose that u u 0 in E and note that
φ ( u ) v φ ( u 0 ) v = R ( W u ( t , u ( t ) ) W u ( t , u 0 ( t ) ) , v ( t ) ) d t .
Combining Lemma 2.2 and the Hölder inequality, we have
φ ( u ) v φ ( u 0 ) v 0 as  u u 0 .
So φ is continuous and Ψ C 1 ( E , R ) . Let u k u in E, we get
φ ( u k ) φ ( u ) E = sup v = 1 ( φ ( u k ) φ ( u ) ) v = sup v = 1 | R ( W u ( t , u k ( t ) ) W u ( t , u ( t ) ) , v ( t ) ) d t | sup v = 1 ( R | W u ( t , u k ( t ) ) W u ( t , u ( t ) ) | 2 d t ) 1 2 v 2 δ 2 ( R | W u ( t , u k ( t ) ) W u ( t , u ( t ) ) | 2 d t ) 1 2 0

as k . Consequently, φ is weakly continuous. Therefore, φ is compact by the weakly continuity of φ since E is a Hilbert space.

Finally, as in the discussion in Lemma 3.1 of [29], we find that the critical points of Ψ are classical solutions of system (1.1) satisfying u C 2 ( R , R N ) , u ( t ) 0 and u ˙ ( t ) 0 as | t | . The proof is complete. □

Remark 2.2 If condition (C4) is replaced by conditions (C7) and (C8), then we can obtain the same conclusion.

In the next section we shall prove our results applying the dual fountain theorem obtained in [35] (see also Proposition 2.1 of [24]). Assume that E be a Banach space with the norm and E = j N X j ¯ , where X j is a finite-dimensional subspace of E. For each k N , let Y k = j = 0 k X j , Z k = j = k X j ¯ . The functional Ψ is said to satisfy the ( P S ) condition if for any sequence { u j } for which { Ψ ( u j ) } is bounded, u j Y k j for some k j with k j and ( Ψ | Y k j ) ( u j ) 0 as j has a subsequence converging to a critical point of Ψ.

Theorem 2.1 Suppose that the functional Ψ C 1 ( E , R ) is even and satisfies the ( P S ) condition. Assume that for each sufficiently large k N , there exist ρ k > r k > 0 such that

(H1) a k : = inf u Z k , u = ρ k Ψ ( u ) 0 .

(H2) b k : = max u Y k , u = r k Ψ ( u ) < 0 .

(H3) d k : = inf u Z k , u ρ k Ψ ( u ) 0 as k .

Then Ψ has a sequence of negative critical values converging to 0.

3 Proof of Theorem 1.3 and 1.4

Now we give the proof of Theorem 1.3.

Proof of Theorem 1.3 We choose a completely orthonormal basis { e j } of X and define X j : = R e j , then Z k and Y k can be defined as that in Section 2. By (C2) and Lemma 2.3, we see that Ψ C 1 ( E , R ) is even. In the following, we will check that all conditions in Theorem 2.1 are satisfied.

Step 1. We prove that Ψ satisfies the ( P S ) condition. Let { u j } be a ( P S ) sequence, that is, { Ψ ( u j ) } is bounded, u j Y k j for some k j with k j and ( Ψ | Y k j ) ( u j ) 0 as j . Now we show that { u j } is bounded in E. By virtue of (C2), (C3) and (C5), for j large enough, we have
ρ M 2 + M 3 u j ρ Ψ ( u j ) Ψ ( u j ) u j = ( ρ 2 1 ) u j 2 + R [ ( W u ( t , u j ) , u j ) ρ W ( t , u j ) ] d t = ( ρ 2 1 ) u j 2 + R [ ( F u ( t , u j ) , u j ) ρ F ( t , u j ) ] d t + R [ ( G u ( t , u j ) , u j ) ρ G ( t , u j ) ] d t ( ρ 2 1 ) u j 2 ( ρ + 1 ) R ( c 2 ( t ) | u j ( t ) | γ + c 3 | u j ( t ) | σ ) d t R h ( t ) | u j ( t ) | δ d t ( ρ 2 1 ) u j 2 ( ρ + 1 ) ( c 2 2 2 γ u j 2 γ + c 3 2 2 σ u j 2 σ ) h 2 2 δ u j 2 δ ( ρ 2 1 ) u j 2 ( ρ + 1 ) ( δ 2 γ c 2 2 2 γ u j γ + δ 2 σ c 3 2 2 σ u j σ ) δ 2 δ h 2 2 δ u j δ
(3.1)

for some M 2 > 0 , M 3 > 0 . Since ρ > 2 and γ , σ , δ < 2 , it follows that { u j } is bounded in E.

From the reflexivity of E, we may extract a weakly convergent subsequence, which, for simplicity, we call { u j } , u j u 1 in E. In view of the Riesz representation theorem, ( Ψ | Y k j ) : Y k j Y k j and φ : E E can be viewed as ( Ψ | Y k j ) : Y k j Y k j and φ : E E , respectively, where Y k j is the dual space of Y k j . Note that
( Ψ | Y k j ) ( u j ) = u j P k j φ ( u j ) , j N ,
(3.2)
where P k j : E Y k j is the orthogonal projection for all j N . That is,
u j = ( Ψ | Y k j ) ( u j ) + P k j φ ( u j ) , j N .
(3.3)

Due to the compactness of φ and u j u 1 , the right hand side of (3.3) converges strongly in E and hence u j u 1 in E.

Step 2. We verify condition (H1) in Theorem 2.1. Set β k = sup u Z k , u = 1 u 2 , then β k 0 as k since E is compactly embedded into L 2 ( R , R N ) . By (C2), (C3) and (C4), we have
Ψ ( u ) = 1 2 u 2 R W ( t , u ( t ) ) d t 1 2 u 2 R ( c 2 ( t ) | u ( t ) | γ + c 3 ( t ) | u ( t ) | σ ) d t R ( a ( t ) | u ( t ) | + c ( t ) | u ( t ) | μ ) d t 1 2 u 2 c 2 2 2 γ u 2 γ c 3 2 2 σ u 2 σ a 2 u 2 c u μ μ 1 2 u 2 β k γ c 2 2 2 γ u γ β k σ c 3 2 2 σ u σ β k a 2 u δ μ μ c u μ .
(3.4)
In view of (3.4), μ > 2 and γ , σ > 1 , one has
Ψ ( u ) 1 4 u 2 ( β k γ c 2 2 2 γ + β k σ c 3 2 2 σ + β k a 2 ) u
(3.5)
for u small enough. Let ρ k = 8 ( β k γ c 2 2 2 γ + β k σ c 3 2 2 σ + β k a 2 ) , it is easy to see that ρ k 0 as k . Thus, for each sufficiently large k N , by (3.5), we get
a k 1 8 ρ k 2 > 0 .
Step 3. We verify condition (H3) in Theorem 2.1. By (3.5), for any u Z k with u ρ k , we have
Ψ ( u ) ( β k γ c 2 2 2 γ + β k σ c 3 2 2 σ + β k a 2 ) u .
(3.6)
Therefore, by (C2), we obtain
0 d k ( β k γ c 2 2 2 γ + β k σ c 3 2 2 σ + β k a 2 ) u .
(3.7)
Since β k , ρ k 0 as k , one has
d k = inf u Z k , u ρ k Ψ ( u ) 0 as  k .
Step 4. We verify condition (H2) in Theorem 2.1. Firstly, we claim that there exists ε > 0 such that
meas { t R : c 1 ( t ) | u ( t ) | γ ε u γ } ε , u Y k { 0 } .
(3.8)
If not, there exists a sequence { u n } Y k with u n = 1 such that
meas { t R : c 1 ( t ) | u n ( t ) | γ 1 n } 1 n .
(3.9)
Since dim Y k < , it follows from the compactness of the unit sphere of Y k that there exists a subsequence, say { u n } , such that u n converges to some u 0 in Y k . Hence, we have u 0 = 1 . Since all norms are equivalent in the finite-dimensional space, we have u n u 0 in L 2 ( R , R N ) . By the Hölder inequality, one has
R c 1 ( t ) | u n u 0 | γ d t c 1 2 2 γ ( R | u n u 0 | 2 d t ) γ 2 0 as  n .
(3.10)
Thus there exist ϵ 1 , ϵ 2 > 0 such that
meas { t R : c 1 ( t ) | u 0 ( t ) | γ ϵ 1 } ϵ 2 .
(3.11)
In fact, if not, we have
meas { t R : c 1 ( t ) | u 0 ( t ) | γ 1 n } = 0 ,
(3.12)
for all positive integers n, which implies that
R c 1 ( t ) | u 0 ( t ) | γ + 2 d t < 1 n u 0 2 2 δ 2 2 n u 0 2 = δ 2 2 n 0
as n . Hence u 0 = 0 , which contradicts u 0 = 1 . Therefore, (3.11) holds. Thus, define
Ω 0 = { t R : c 1 ( t ) | u 0 ( t ) | γ ϵ 1 } , Ω n = { t R : c 1 ( t ) | u n ( t ) | γ < 1 n }
and Ω n c = R Ω n = { t R : c 1 ( t ) | u n ( t ) | γ 1 n } . Combining (3.9) and (3.11), we have
meas ( Ω n Ω 0 ) = meas ( Ω 0 Ω n c Ω 0 ) meas ( Ω 0 ) meas ( Ω n c Ω 0 ) ϵ 2 1 n
for all positive integers n. Let n be large enough such that ϵ 2 1 n 1 2 ϵ 2 and 1 2 γ 1 ϵ 1 1 n 1 2 γ ϵ 1 . Then we have
R c 1 ( t ) | u n u 0 | γ d t Ω n Ω 0 c 1 ( t ) | u n u 0 | γ d t 1 2 γ 1 Ω n Ω 0 c 1 ( t ) | u 0 | γ d t Ω n Ω 0 c 1 ( t ) | u n | γ d t ( 1 2 γ 1 ϵ 1 1 n ) meas ( Ω n Ω 0 ) ϵ 1 ϵ 2 2 γ + 1
for all large n, which is a contradiction to (3.10). Therefore, (3.8) holds. For the ε given in (3.8), let
Ω u = { t R : c 1 ( t ) | u ( t ) | γ ε u γ } , u Y k { 0 } .
(3.13)
By (3.8), we obtain
meas ( Ω u ) ε , u Y k { 0 } .
(3.14)
For any u k Y k , by (C2), (C3), (C4), (3.13) and (3.14), we have
Ψ ( u ) = 1 2 u 2 R W ( t , u ( t ) ) d t 1 2 u 2 1 γ R c 1 ( t ) | u ( t ) | γ d t R G ( t , u ( t ) ) d t 1 2 u 2 1 γ R c 1 ( t ) | u ( t ) | γ d t 1 2 u 2 1 γ Ω u c 1 ( t ) | u ( t ) | γ d t 1 2 u 2 ε γ u γ meas ( Ω u ) 1 2 u 2 ε 2 γ u γ .
Choose 0 < r k < min { ρ k , ( ε 2 / γ ) 1 2 γ } . Direct computation shows that
b k r k 2 2 < 0 , k N .

Thus, by Theorem 2.1, Ψ has infinitely many nontrivial critical points, that is, system (1.1) possesses infinitely many homoclinic solutions. □

Now we give the proof of Theorem 1.4.

Proof of Theorem 1.4 Step 1. We prove that Ψ satisfies the ( P S ) condition. Let { u j } be a ( P S ) sequence, that is, { Ψ ( u j ) } is bounded, u j Y k j for some k j with k j and ( Ψ | Y k j ) ( u j ) 0 as j . Now we show that { u j } is bounded in E. In view of (C1), (C2), (C3) and (C6), for j large enough, we obtain
λ M 4 + M 5 u j λ Ψ ( u j ) Ψ ( u j ) u j = ( λ 2 1 ) u j 2 + R [ ( W u ( t , u j ) , u j ) λ W ( t , u j ) ] d t = ( λ 2 1 ) u j 2 + R [ ( F u ( t , u j ) , u j ) λ F ( t , u j ) ] d t + R [ ( G u ( t , u j ) , u j ) λ G ( t , u j ) ] d t ( λ 2 1 ) u j 2 ( λ + 1 ) R ( c 2 ( t ) | u j ( t ) | γ + c 3 | u j ( t ) | σ ) d t h 1 R | u j ( t ) | 2 d t ( λ 2 1 ) u j 2 ( λ + 1 ) ( c 2 2 2 γ u j 2 γ + c 3 2 2 σ u j 2 σ ) h 1 β 0 R ( L ( t ) u j ( t ) , u j ( t ) ) d t ( λ 2 2 h 1 β 0 ) u j 2 ( λ + 1 ) ( δ 2 γ c 2 2 2 γ u j γ + δ 2 σ c 3 2 2 σ u j σ )
(3.15)

for some M 4 > 0 , M 5 > 0 . Since 0 h 1 < β 0 ( λ 2 ) 2 and 1 < γ , σ < 2 , it follows that { u j } is bounded in E. In the following, the proof of the ( P S ) condition is the same as that in Theorem 1.1, and we omit it here.

Step 2. We verify condition (H1) in Theorem 2.1. Set β k = sup u Z k , u = 1 u 2 , then β k 0 as k since E is compactly embedded into L 2 ( R , R N ) . By (C8), we have
| G ( t , u ) | h 2 ( | u | + | u | p 1 ) .
(3.16)
It follows from (C7) that there exists δ 0 > 0 such that
| G ( t , u ) | θ 2 | u | 2 , | u | δ 0 , t R .
(3.17)
Combining (3.16) and (3.17), we get
| G ( t , u ) | θ 2 | u | 2 + h 3 | u | p 1 , ( t , u ) R × R N ,
(3.18)
where h 3 = h 2 ( 1 + δ 0 1 p 1 ) . By (C1), (C2), (C3) and (3.18), we have
Ψ ( u ) = 1 2 u 2 R W ( t , u ( t ) ) d t 1 2 u 2 R ( c 2 ( t ) | u ( t ) | γ + c 3 ( t ) | u ( t ) | σ ) d t R ( θ 2 | u ( t ) | 2 + h 3 | u ( t ) | p 1 ) d t 1 2 u 2 c 2 2 2 γ u 2 γ c 3 2 2 σ u 2 σ h 3 u p 1 p 1 θ 2 β 0 R ( L ( t ) u j ( t ) , u j ( t ) ) d t ( 1 2 θ 2 β 0 ) u 2 β k γ c 2 2 2 γ u γ β k σ c 3 2 2 σ u σ h 3 δ p 1 p 1 u p 1 .
(3.19)
Take θ 0 = 1 2 θ 2 β 0 , by (C7), we obtain θ 0 > 0 . By virtue of (3.19), p 1 > 2 and γ , σ > 1 , one has
Ψ ( u ) θ 0 2 u 2 ( β k γ c 2 2 2 γ + β k σ c 3 2 2 σ ) u
(3.20)
for u small enough. Let ρ k = θ 0 4 ( β k γ c 2 2 2 γ + β k σ c 3 2 2 σ ) , it is easy to see that ρ k 0 as k . Thus, for each sufficiently large k N , by (3.20), we get
a k θ 0 4 ρ k 2 > 0 .

Step 3. We verify condition (H3) in Theorem 2.1. The proof is similar to the Step 3 in the proof of Theorem 1.3, and we omit it.

Step 4. We verify condition (H2) in Theorem 2.1. The proof is the same as that the Step 4 in the proof of Theorem 1.3, and we omit it here.

Thus, by Theorem 2.1, Ψ has infinitely many nontrivial critical points, that is, system (1.1) possesses infinitely many homoclinic solutions. □

Declarations

Acknowledgements

The authors would like to thank the referees for their valuable suggestions and comments which lead to improvement of the manuscript.

Authors’ Affiliations

(1)
School of Mathematics and Statistics, Central South University

References

  1. Mawhin J, Willem M: Critical Point Theory and Hamiltonian Systems. Springer, Berlin; 1989.View ArticleGoogle Scholar
  2. Rabinowitz PH: Variational methods for Hamiltonian systems. 1. In Handbook of Dynamical Systems. North-Holland, Amsterdam; 2002:1091–1127. part 1, Chapter 14Google Scholar
  3. Bonanno G, Livrea R: Multiple periodic solutions for Hamiltonian systems with not coercive potential. J. Math. Anal. Appl. 2010, 363: 627–638. 10.1016/j.jmaa.2009.09.025MathSciNetView ArticleGoogle Scholar
  4. Caldiroli P, Montecchiari P: Homoclinic orbits for second order Hamiltonian systems with potential changing sign. Commun. Appl. Nonlinear Anal. 1994, 1(2):97–129.MathSciNetGoogle Scholar
  5. Chen P, Tang XH, Agarwal RP:Infinitely many homoclinic solutions for nonautonomous p ( t ) -Laplacian Hamiltonian systems. Comput. Math. Appl. 2012, 63: 751–763. 10.1016/j.camwa.2011.11.039MathSciNetView ArticleGoogle Scholar
  6. Cordaro G, Rao G: Three periodic solutions for perturbed second order Hamiltonian systems. J. Math. Anal. Appl. 2009, 359: 780–785. 10.1016/j.jmaa.2009.06.049MathSciNetView ArticleGoogle Scholar
  7. Coti Zelati V, Rabinowitz PH: Homoclinic orbits for second order Hamiltonian systems possessing superquadratic potentials. J. Am. Math. Soc. 1991, 4: 693–727. 10.1090/S0894-0347-1991-1119200-3MathSciNetView ArticleGoogle Scholar
  8. Ding Y: Existence and multiplicity results for homoclinic solutions to a class of Hamiltonian systems. Nonlinear Anal. 1995, 25: 1095–1113. 10.1016/0362-546X(94)00229-BMathSciNetView ArticleGoogle Scholar
  9. Ding Y, Lee C: Homoclinics for asymptotically quadratic and superquadratic Hamiltonian systems. Nonlinear Anal. 2009, 71: 1395–1413. 10.1016/j.na.2008.10.116MathSciNetView ArticleGoogle Scholar
  10. Guo C, Agarwal RP, Wang C, O’Regan D: The existence of homoclinic orbits for a class of first-order superquadratic Hamiltonian systems. Mem. Differ. Equ. Math. Phys. 2014, 61: 83–102.MathSciNetGoogle Scholar
  11. Korman P, Lazer AC: Homoclinic orbits for a class of symmetric Hamiltonian systems. Electron. J. Differ. Equ. 1994, 1994: 1–10.MathSciNetGoogle Scholar
  12. Lv X, Lu S, Jiang J: Homoclinic solutions for a class of second-order Hamiltonian systems. Nonlinear Anal., Real World Appl. 2012, 13: 176–185. 10.1016/j.nonrwa.2011.07.023MathSciNetView ArticleGoogle Scholar
  13. Omana W, Willem M: Homoclinic orbits for a class of Hamiltonian systems. Differ. Integral Equ. 1992, 5: 1115–1120.MathSciNetGoogle Scholar
  14. Paturel E: Multiple homoclinic orbits for a class of Hamiltonian systems. Calc. Var. Partial Differ. Equ. 2001, 12: 117–143. 10.1007/PL00009909MathSciNetView ArticleGoogle Scholar
  15. Qu Z, Tang C: Existence of homoclinic solutions for the second order Hamiltonian systems. J. Math. Anal. Appl. 2004, 291: 203–213. 10.1016/j.jmaa.2003.10.026MathSciNetView ArticleGoogle Scholar
  16. Rabinowitz PH: Homoclinic orbits for a class of Hamiltonian systems. Proc. R. Soc. Edinb., Sect. A 1990, 114: 33–38. 10.1017/S0308210500024240MathSciNetView ArticleGoogle Scholar
  17. Rabinowitz PH, Tanaka K: Some results on connecting orbits for a class of Hamiltonian systems. Math. Z. 1991, 206: 473–499. 10.1007/BF02571356MathSciNetView ArticleGoogle Scholar
  18. Sun J, Chen H, Nieto JJ: Homoclinic solutions for a class of subquadratic second-order Hamiltonian systems. J. Math. Anal. Appl. 2011, 373: 20–29. 10.1016/j.jmaa.2010.06.038MathSciNetView ArticleGoogle Scholar
  19. Tang X, Lin X: Existence of infinitely many homoclinic orbits in Hamiltonian systems. Proc. R. Soc. Edinb., Sect. A 2011, 141: 1103–1119. 10.1017/S0308210509001346MathSciNetView ArticleGoogle Scholar
  20. Tang X, Lin X: Homoclinic solutions for a class of second-order Hamiltonian systems. J. Math. Anal. Appl. 2009, 354: 539–549. 10.1016/j.jmaa.2008.12.052MathSciNetView ArticleGoogle Scholar
  21. Wang J, Xu J, Zhang F: Homoclinic orbits for a class of Hamiltonian systems with superquadratic or asymptotically quadratic potentials. Commun. Pure Appl. Anal. 2011, 10: 269–286.MathSciNetView ArticleGoogle Scholar
  22. Wei J, Wang J: Infinitely many homoclinic orbits for second order Hamiltonian systems with general potentials. J. Math. Anal. Appl. 2010, 366: 694–699. 10.1016/j.jmaa.2009.12.024MathSciNetView ArticleGoogle Scholar
  23. Yang L, Chen H, Sun J: Infinitely many homoclinic solutions for some second order Hamiltonian systems. Nonlinear Anal. 2011, 74: 6459–6468. 10.1016/j.na.2011.06.029MathSciNetView ArticleGoogle Scholar
  24. Yang M, Han Z: Infinitely many homoclinic solutions for second-order Hamiltonian systems with odd nonlinearities. Nonlinear Anal. 2011, 74: 2635–2646. 10.1016/j.na.2010.12.019MathSciNetView ArticleGoogle Scholar
  25. Yang J, Zhang F: Infinitely many homoclinic orbits for the second-order Hamiltonian systems with super-quadratic potentials. Nonlinear Anal., Real World Appl. 2009, 10: 1417–1423. 10.1016/j.nonrwa.2008.01.013MathSciNetView ArticleGoogle Scholar
  26. Yu JS: Subharmonic solutions with prescribed minimal period of a class of nonautonomous Hamiltonian systems. J. Dyn. Differ. Equ. 2008, 20: 787–796. 10.1007/s10884-008-9118-1View ArticleGoogle Scholar
  27. Zhang Q, Liu C: Infinitely many homoclinic solutions for second order Hamiltonian systems. Nonlinear Anal. 2010, 72: 894–903. 10.1016/j.na.2009.07.021MathSciNetView ArticleGoogle Scholar
  28. Zhang Z, Yuan R: Homoclinic solutions of some second order non-autonomous systems. Nonlinear Anal. 2009, 71: 5790–5798. 10.1016/j.na.2009.05.003MathSciNetView ArticleGoogle Scholar
  29. Zhang Z, Yuan R: Homoclinic solutions for a class of non-autonomous subquadratic second-order Hamiltonian systems. Nonlinear Anal. 2009, 71: 4125–4130. 10.1016/j.na.2009.02.071MathSciNetView ArticleGoogle Scholar
  30. Zou W, Li S: Infinitely many homoclinic orbits for the second-order Hamiltonian systems. Appl. Math. Lett. 2003, 16: 1283–1287. 10.1016/S0893-9659(03)90130-3MathSciNetView ArticleGoogle Scholar
  31. Ricceri B: A three critical points theorem revisited. Nonlinear Anal. 2009, 70: 3084–3089. 10.1016/j.na.2008.04.010MathSciNetView ArticleGoogle Scholar
  32. Ricceri B: A further three critical points theorem. Nonlinear Anal. 2009, 71: 4151–4157. 10.1016/j.na.2009.02.074MathSciNetView ArticleGoogle Scholar
  33. Chen HW, He ZM: New results for perturbed Hamiltonian systems with impulses. Appl. Math. Comput. 2012, 218: 9489–9497. 10.1016/j.amc.2012.03.038MathSciNetView ArticleGoogle Scholar
  34. Sun J, Chen H, Nieto JJ, Otero-Novoa M: The multiplicity of solutions for perturbed second-order Hamiltonian systems with impulsive effects. Nonlinear Anal. 2010, 72: 4575–4586. 10.1016/j.na.2010.02.034MathSciNetView ArticleGoogle Scholar
  35. Bartsch T, Willem M: On an elliptic equation with concave and convex nonlinearities. Proc. Am. Math. Soc. 1995, 123: 3555–3561. 10.1090/S0002-9939-1995-1301008-2MathSciNetView ArticleGoogle Scholar

Copyright

© Chen and He; licensee Springer. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.