AAA Abstract and Applied Analysis 1687-0409 1085-3375 Hindawi Publishing Corporation 359105 10.1155/2014/359105 359105 Research Article Nontrivial Periodic Solutions to Some Semilinear Sixth-Order Difference Equations Long Yuhua 1, 2 Wang Youyu 1 School of Mathematics and Information Science Guangzhou University Guangzhou 510006 China gzhu.edu.cn 2 Key Laboratory of Mathematics and Interdisciplinary Sciences of Guangdong Higher Education Institute Guangzhou University Guangzhou 510006 China gzhu.edu.cn 2014 1232014 2014 18 11 2013 23 01 2014 12 3 2014 2014 Copyright © 2014 Yuhua Long. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

We establish some new criteria to guarantee nonexistence, existence, and multiplicity of nontrivial periodic solutions of some semilinear sixth-order difference equations by using minmax method, Z 2 index theory, and variational technique. Our results only make some assumptions on the period T , which are very easy to verify and rather relaxed.

1. Introduction

In the present paper, we deal with the following sixth-order difference equation: (1) Δ 6 u t - 3 + A Δ 4 u t - 2 + B Δ 2 u t - 1 + u t - u t 3 = 0 , t Z with u 0 = u T = 0 , where T 2 is an integer and [ 1 , T ] denotes the discrete interval { 1,2 , , T } . Δ is the forward difference operator defined by Δ u t = u t + 1 - u t and Δ n u t - k = Δ n - 1 ( Δ u t - k ) . A and B are positive constants satisfying A 2 < 4 B .

By using Z 2 index theory in combination with variational technique, we will prove nonexistence, existence, and multiplicity of nontrivial periodic solutions to (1) under convenient assumptions on T . All our results only depend on A and B and are easy to satisfy.

Periodic solution problems for difference equations have been extensively studied (see the monographs of Lakshmikantham and Trigiante  and of Agarwal ). The classical theory of difference equations employs numerical analysis and features from the linear and nonlinear operator theory, such as fixed point methods; we remark that, usually, the applications of fixed point methods yield existence results only. Recently, although many new results have been established by applying variational methods, we recall here the works of Cai and Yu , Guo and Yu , and Deng et al. . The variational approach represents an important advance as it allows proving multiplicity results as well.

In general, (1) may be regarded as a discrete analogue of the following sixth-order differential equation: (2) u t = 6 u x 6 + A 4 u x 4 + B 2 u x 2 + u - u 3 . As to (2), it is a model for describing the behavior of phase fronts in materials that are undergoing a transition between the liquid and solid state and be widely studied; one can refer to  and references therein.

Difference equations, the discrete analogs of differential equations, represent the discrete counterpart of ordinary differential equations and are usually studied in connection with numerical analysis. They occur widely in numerous settings and forms, both in mathematics itself and in its applications to computing, statistics, electrical circuit analysis, biology, dynamical systems, economics, and other fields. For the general background of difference equations, one can refer to monographs [2, 913] for details.

Since 2003, critical point theory has been a powerful tool to establish sufficient conditions on the existence of periodic solutions of difference equations and many significant results have been obtained; see, for example, [4, 14, 15]. Compared to first-order or second-order difference equations, the study of higher order difference equations has received considerably less attention. For example,  studied (3) i = 0 n Δ i ( r i ( k - i ) Δ i u ( k - i ) ) = 0 in the context of discrete calculus of variational functional, and Peil and Peterson  studied the asymptotic behaviour of solutions of (3) with r i ( k ) 0 for 1 i n - 1 . In 2007, based on Linking Theorem,  gave some criteria for the existence of periodic solutions of (4) Δ n ( r t - n ) Δ n u t - n + f ( t , u t ) = 0 ,    bbbbbbbbb n [ 1,3 ] , n Z for the case where f grows superlinear at both 0 and . Results in  made many assumptions on f and they are not easy to verify. The aim of this paper is to apply critical point theory to deal with the periodic solution problems of (1) when it is semilinear under concise and explicit assumptions on the period T . The main results of this paper are the following three theorems.

Theorem 1.

Let A > 0 , A 2 < 4 B , and integer T satisfy (5) 0 < T < T 1 , where T 1 is defined as (6) sin π T 1 = max { ( 1 4 B - A 2 ) 1 / 2 , 1 2 ( 4 B 4 B - A 2 ) 1 / 6 } then (1) has only the trivial solution.

Theorem 2.

Let A > 0 , A 2 < 4 B , and integer T satisfy (7) T > T 2 , where T 2 is defined as (8) sin π 2 T 2 = 1 2 ( 1 2 + 1 4 - B 3 27 3 + 1 2 - 1 4 - B 3 27 3 ) 1 / 2 ; there exists a nontrivial solution of (1).

Theorem 3.

Let A > 0 , A 2 < 4 B , and (9) T > m T 2 for some m N ; there exist at least m geometrically distinct nontrivial solutions of (1).

The remaining of the paper is organized as follows. In Section 2 we establish the variational framework associated with (1) and transfer the problem on the existence of periodic solutions of (1) into the existence of critical points of the corresponding functional. We also state some fundamental lemmas for later use. Then we present the detailed proofs of main results in Section 3. Finally, we exhibit a simple example to illustrate our conclusions.

2. Preliminaries

In order to study the periodic solutions of (1), we will state some basic notations and lemmas, which will be used in the proofs of our main results. Let (10) S = { u = { u t } u t R N , t Z } . For a given integer T 2 , E T is defined as a subspace of S by (11) E T = { u = { u t } S u t + T = u t , t Z } , and for u , v E T , let (12) u , v = t = 1 T ( Δ 3 u t - 1 Δ 3 v t - 1 + Δ 2 u t - 1 Δ 2 v t - 1 ssssssss + Δ u t - 1 Δ v t - 1 + u t v t u t - 1 Δ 3 ) . Then E T is a finite dimensional Hilbert space with above inner product, and the induced norm is (13) u = ( t = 1 T ( ( Δ 3 u t - 1 ) 2 + ( Δ 2 u t - 1 ) 2 + ( Δ u t - 1 ) 2 + u t 2 ) ) 1 / 2 ssssssssssssssssssssssssssssssssssssssssssssssss u E T . As usual, for 1 p < + , let (14) l p ( Z , R N ) = { u S : t Z | u t | p < + } and its norm is defined by (15) u l p = ( t = 1 T | u t | p ) 1 / p , u l p ( Z , R N ) .

Define the functional I : E T R as follows: (16) I ( u ; T ) = t = 1 T [ 1 2 ( ( Δ 3 u t - 1 ) 2 - A ( Δ 2 u t - 1 ) 2 ssssssssssss + B ( Δ u t - 1 ) 2 - u t 2 ( Δ 2 u t - 1 ) 2 ) + 1 4 u t 4 ] . Clearly, I ( u ; T ) C 1 ( E T , R ) and for any u , v E T one can easily check that (17) I ( u ; T ) , v = t = 1 T ( Δ 3 u t - 1 Δ 3 v t - 1 - A Δ 2 u t - 1 Δ 2 v t - 1 ssssssss + B Δ u t - 1 Δ v t - 1 - u t v t + u t 3 v t ) . For any u = { u t } t Z E T , by using u i = u T + i for any i Z , (18) Δ n u t - 1 = k = 0 n ( - 1 ) k ( n k ) u t + n - k - 1 . We can compute the partial derivative as (19) I u t = Δ 6 u t - 3 + A Δ 4 u t - 2 + B Δ 2 u t - 1 + u t - u t 3 , t [ 1 , T ] . Then, u = { u t } t Z is a critical point of I ( u ; T ) on E T ; that is, I ( u ; T ) = 0 if and only if (20) Δ 6 u t - 3 + A Δ 4 u t - 2 + B Δ 2 u t - 1 + u t - u t 3 = 0 , t [ 1 , T ] . By the periodicity of u t , we have reduced the existence of periodic solutions of (1) to the existence of critical points of I ( u ; T ) on E T . For convenience, we identity u E T with u = ( u 1 , u 2 , , u T ) T , so we draw a conclusion as follows.

Lemma 4.

Suppose that u t is a critical point of the functional I ( u ; T ) ; then u t is a T -periodic solution of (1).

We provide some lemmas which will be needed in proofs of our main results.

Lemma 5.

For any x ( j ) > 0 , y ( j ) > 0 , j [ 1 , n ] , and n Z , (21) j = 1 n x ( j ) y ( j ) ( j = 1 n x r ( j ) ) 1 / r · ( j = 1 n y s ( j ) ) 1 / s , where r > 1 , s > 1 , and ( 1 / r ) + ( 1 / s ) = 1 .

Lemma 6.

Let u E T be a critical point of I ( u ; T ) ; for every v E T , there hold (22) t = 1 T Δ u t - 1 Δ v t - 1 = - t = 1 T Δ 2 u t - 1 v t , t = 1 T Δ 2 u t - 1 Δ 2 v t - 1 = - t = 1 T Δ u t - 1 Δ 3 v t - 1 .

Proof.

Let u E T be a critical point of I ( u ; T ) , according to the definition of Δ and the periodicity of u t and v t ; then we have (23) t = 1 T Δ u t - 1 Δ v t - 1 = t = 1 T ( u t - u t - 1 ) ( v t - v t - 1 ) = t = 1 T ( u t v t - u t - 1 v t - 1 ) - t = 1 T ( u t v t - 1 - u t - 1 v t - 1 ) = t = 1 T Δ u t - 1 v t - t = 1 T ( u t + 1 v t - u t v t ) = t = 1 T ( Δ u t - 1 - Δ u t ) v t = - t = 1 T Δ 2 u t - 1 v t . Similarly, we get (24) t = 1 T Δ 2 u t - 1 Δ 2 v t - 1 = - t = 1 T Δ u t - 1 Δ 3 v t - 1 .

Let X be a real Banach space. Σ is the subset of X , which is closed and symmetric with respect to 0; that is, (25) Σ = { A X A    is    closed    and    x A    iff    - x A } . For any A Σ , the Z 2 geometric index, also called genus, γ of A is defined by (26) γ ( A ) = min { k N there    is    an    odd    map    φ C ( X , R k { 0 } ) } , when there exists no such finite k , set γ ( A ) = . Finally set γ ( ) = 0 .

The following lemma is trivial.

Lemma 7 (Chang [<xref ref-type="bibr" rid="B18">18</xref>] and Rabinowitz [<xref ref-type="bibr" rid="B19">19</xref>]).

Let X k be a subset of X with dimension k and S 1 the unit sphere of X and then let γ ( X k S 1 ) = k .

Next, let us recall the definition of Palais-Smale condition.

Let X be a real Banach space, Φ C ( X , R ) . Φ is a continuously Frechet differentiable functional defined on X . Φ is said to be satisfied Palais-Smale condition (P.S. condition for short) if any sequence { u n } X for which { Φ ( u n ) } is bounded and Φ ( u n ) 0 ( n ) possesses a convergent subsequence in X .

Lemma 8 (see [<xref ref-type="bibr" rid="B20">20</xref>]).

Let E be a real Banach space and let f : E R be a C 1 functional and satisfy P.S. condition. If f is bound from below, then (27) c = inf u E f ( u ) is a critical value of f .

Lemma 9 (see [<xref ref-type="bibr" rid="B19">19</xref>]).

Let Φ be an even functional defined on X and satisfy P.S. condition. For positive integer j , define (28) c j = sup γ ( A ) j A Σ inf u A Φ ( u ) . Then (29) - c k c k - 1 c 2 c 1 + and

assume - < c j < + ; then c j is a critical value of Φ ;

if c = c j = c j + 1 = = c j + l - 1 , then γ ( K c ) l , where K c = { u X Φ ( u ) = c , Φ ( u ) = 0 } .

3. Proofs of Main Results

With the above preparations, we will prove our main results in this section. In order to give proofs of theorems, we need the following lemmas.

Lemma 10.

For any u E T , if u t is a critical of I ( u ; T ) , we have (30) t = 1 T u t 2 1 4 sin 2 ( π / T ) t = 1 T ( Δ u t ) 2 , (31) t = 1 T u t 2 1 16 sin 4 ( π / T ) t = 1 T ( Δ 2 u t - 1 ) 2 , (32) t = 1 T u t 2 1 64 sin 6 ( π / T ) t = 1 T ( Δ 3 u t - 1 ) 2 .

Proof.

From Agarwal , we have inequality (30) and (33) t = 1 T ( Δ u t - 1 ) 2 1 4 sin 2 ( π / T ) t = 1 T ( Δ 2 u t - 1 ) 2 ; then (34) t = 1 T u t 2 1 4 sin 2 ( π / T ) · 1 4 sin 2 ( π / T ) t = 1 T ( Δ 2 u t - 1 ) 2 = 1 16 sin 4 ( π / T ) t = 1 T ( Δ 2 u t - 1 ) 2 that is inequality (31).

From Lemmas 5 and 6, there holds (35) t = 1 T ( Δ 2 u t - 1 ) 2 = - t = 1 T Δ u t - 1 · Δ 3 u t - 1 ( t = 1 T ( Δ u t - 1 ) 2 ) 1 / 2 · ( t = 1 T ( Δ 3 u t - 1 ) 2 ) 1 / 2 = ( - t = 1 T u t Δ 2 u t - 1 ) 1 / 2 · ( t = 1 T ( Δ 3 u t - 1 ) 2 ) 1 / 2 ( t = 1 T u t 2 ) 1 / 4 · ( t = 1 T ( Δ 2 u t - 1 ) 2 ) 1 / 4 · ( t = 1 T ( Δ 3 u t - 1 ) 2 ) 1 / 2 and it follows (36) t = 1 T ( Δ 2 u t - 1 ) 2 ( t = 1 T u t 2 ) 1 / 3 · ( t = 1 T ( Δ 3 u t - 1 ) 2 ) 2 / 3 . Using inequality (31), we have inequality (32) is true.

Lemma 11.

Suppose A > 0 and A 2 < 4 B , u E T is a critical point of I ( u ; T ) , and then (37) A t = 1 T ( Δ 2 u t - 1 ) 2 A 2 4 t = 1 T ( Δ u t - 1 ) 2 + t = 1 T ( Δ 3 u t - 1 ) 2 , (38) A t = 1 T ( Δ 2 u t - 1 ) 2 B t = 1 T ( Δ u t - 1 ) 2 + A 2 4 B t = 1 T ( Δ 3 u t - 1 ) 2 .

Proof.

From Lemmas 5 and 6, we have (39) A t = 1 T ( Δ 2 u t - 1 ) 2 = - A t = 1 T Δ u t - 1 Δ 3 u t - 1 = - t = 1 T 2 2 A Δ u t - 1 · 2 Δ 3 u t - 1 ( t = 1 T ( 2 2 A Δ u t - 1 ) 2 ) 1 / 2 · ( t = 1 T 2 ( Δ 3 u t - 1 ) 2 ) 1 / 2 1 2 ( t = 1 T A 2 2 ( Δ u t - 1 ) 2 + t = 1 T 2 ( Δ 3 u t - 1 ) 2 ) = A 2 4 t = 1 T ( Δ u t - 1 ) 2 + t = 1 T ( Δ 3 u t - 1 ) 2 , so inequality (37) holds.

To prove inequality (38), let (40) A = A 2 B · 2 B . Similar to the proof of inequality (37), we get inequality (38).

Now we will give the proof of Theorem 1.

Proof of Theorem <xref ref-type="statement" rid="thm1.1">1</xref>.

Suppose all conditions of Theorem 1 hold. Let u E T be a nontrivial solution of (1); then (41) t = 1 T ( ( Δ 3 u t - 1 ) 2 - A ( Δ 2 u t - 1 ) 2 + B ( Δ u t - 1 ) 2 - u t 2 + u t 4 ) = 0 .

By (37), (30), (5), and (6), we have (42) 0 > - t = 1 T u t 4 = t = 1 T ( ( Δ 3 u t - 1 ) 2 - A ( Δ 2 u t - 1 ) 2 + B ( Δ u t - 1 ) 2 - u t 2 ) t = 1 T ( ( Δ 3 u t - 1 ) 2 - A 2 4 ( Δ 2 u t - 1 ) 2 sssssssss - ( Δ 3 u t - 1 ) 2 + B ( Δ u t - 1 ) 2 - u t 2 A 2 4 ) = t = 1 T ( B - A 2 4 ) ( Δ u t - 1 ) 2 - t = 1 T u t 2 ( ( B - A 2 4 ) · 4 sin 2 π T - 1 ) t = 1 T u t 2 0 which is a contradiction. Moreover, by (38), (31), (5), and (6), we have (43) 0 > - t = 1 T u t 4 = t = 1 T ( ( Δ 3 u t - 1 ) 2 - A ( Δ 2 u t - 1 ) 2 + B ( Δ u t - 1 ) 2 - u t 2 ) t = 1 T ( A 2 4 B ( Δ 3 u t - 1 ) 2 ( Δ 3 u t - 1 ) 2 - B ( Δ 2 u t - 1 ) 2 - A 2 4 B ( Δ 3 u t - 1 ) 2 sssssssssss + B ( Δ u t - 1 ) 2 - u t 2 A 2 4 B ( Δ 3 u t - 1 ) 2 ) = t = 1 T ( ( 1 - A 2 4 B ) ( Δ 3 u t - 1 ) 2 - u t 2 ) ( ( 1 - A 2 4 B ) · 64 sin 6 π T - 1 ) t = 1 T u t 2 0 which is a contradiction. Thus u t should be the trivial solution of (1); in other words, (1) has no nontrivial solution when (5) and (6) hold.

To apply Lemmas 8 and 9 to look for nontrivial solutions for (1), next we prove that I ( u ; T ) satisfies P.S. condition.

Lemma 12.

Let A > 0 and A 2 < 4 B , and then the functional I ( u ; T ) is bounded from below on E T and satisfies P.S. condition.

Proof.

Denote ϵ = 1 - ( A 2 / 4 B ) , and then ϵ > 0 . From (38) and the elementary inequality (44) - 1 2 u 2 + 1 4 u 4 - 1 4 there hold (45) I ( u ; T ) = t = 1 T [ 1 2 ( ( Δ 3 u t - 1 ) 2 - A ( Δ 2 u t - 1 ) 2 + B ( Δ u t - 1 ) 2 - u t 2 ) ssssssssss + 1 4 u t 4 ] = t = 1 T 1 2 ( ( Δ 3 u t - 1 ) 2 - A ( Δ 2 u t - 1 ) 2 + B ( Δ u t - 1 ) 2 ) + t = 1 T ( 1 2 u t 2 + 1 4 u t 4 ) t = 1 T ϵ 2 ( Δ 3 u t - 1 ) 2 - t = 1 T 1 4 - T 4 which means that the functional I ( u ; T ) is bounded from below on E T .

Next we will show that I ( u ; T ) satisfies P.S. condition. Suppose { u ( k ) } E T satisfy that { I ( u ( k ) ; T ) } is a bounded sequence from above; that is, there exists a positive constant C such that (46) I ( u ( k ) ; T ) C , k N . By (38), we have (47) ϵ 2 t = 1 T ( Δ 3 u t - 1 ( k ) ) 2 t = 1 T ( ϵ 2 ( Δ 3 u t - 1 ( k ) ) 2 + 1 4 ( ( u t ( k ) ) 2 - 1 ) 2 ) = t = 1 T ( 1 2 ( 1 - A 2 4 B ) ( Δ 3 u t - 1 ( k ) ) 2 + 1 4 ( u t ( k ) ) 4 ssssssssssss - 1 2 ( u t ( k ) ) 2 + 1 4 1 2 ( 1 - A 2 4 B ) ) t = 1 T [ 1 2 ( ( Δ 3 u t - 1 ( k ) ) 2 - A ( Δ 2 u t - 1 ( k ) ) 2 sssssssssssssss + B ( Δ u t - 1 ( k ) ) 2 - ( u t ( k ) ) 2 ) ssssssssss + 1 4 ( u t ( k ) ) 4 + 1 4 ] = I ( u ( k ) ; T ) + T 4 C + T 4 and it follows (48) t = 1 T ( Δ 3 u t - 1 ( k ) ) 2 2 ϵ ( C + T 4 ) . And together with (32), there holds (49) t = 1 T ( u t ( k ) ) 2 2 64 ϵ sin 6 ( π / T ) ( C + T 4 ) . From (37), it follows (50) ϵ 2 t = 1 T ( Δ 2 u t - 1 ( k ) ) 2 t = 1 T ( ϵ 2 ( Δ 2 u t - 1 ( k ) ) 2 + 1 4 ( ( u t ( k ) ) 2 - 1 ) 2 ) t = 1 T [ 1 2 ( ( Δ 2 u t - 1 ( k ) ) 2 - A ( Δ 2 u t - 1 ( k ) ) 2 sssssssssssssss + B ( Δ u t - 1 ( k ) ) 2 - ( u t ( k ) ) 2 ) sssssssssss + 1 4 ( u t ( k ) ) 4 + 1 4 ] = I ( u ( k ) ; T ) + T 4 C + T 4 and then (51) t = 1 T ( Δ 2 u t - 1 ( k ) ) 2 8 4 B - A 2 ( C + T 4 ) ; thus (52) t = 1 T ( Δ u t - 1 ( k ) ) 2 1 4 sin 2 ( π / T ) t = 1 T ( Δ 2 u t - 1 ( k ) ) 2 8 4 ( 4 B - A 2 ) sin 2 ( π / T ) ( C + T 4 ) . Therefore, by (48)–(52), we get that there exists a positive constant M such that (53) t = 1 T ( ( Δ 3 u t - 1 ( k ) ) 2 + ( Δ 2 u t - 1 ( k ) ) 2 + ( Δ u t - 1 ( k ) ) 2 + ( u t ( k ) ) 2 ) M 2 ; that is, (54) u ( k ) M . As a consequence, { u ( k ) } possesses a convergence subsequence in the finite dimensional Hilbert space E T and I ( u ; T ) satisfies P.S. condition. This completes the proof of Lemma 12.

From Lemmas 8 and 12, there exists c = inf u E T I ( u ; T ) is a critical value of I ( u ; T ) , which means that there exists a critical point of I ( u ; T ) on E T . Next, we devote ourselves to verifying the critical point is nontrivial.

Set (55) M ( T ) = min { I ( u ; T ) u E T } . In Theorem 1, we have shown that M ( T ) = 0 if T T 1 . To complete the proof of Theorem 2, we will show this does not hold for large T . We prove that for T > T 2 , where T 2 is an appropriate number defined as (8), it holds M ( T ) < 0 and the corresponding critical point is nontrivial.

Proof of Theorem <xref ref-type="statement" rid="thm1.2">2</xref>.

By Lemma 12, c = inf u E T I ( u ; T ) is nontrivial. So if we can find some critical point u ¯ such that I ( u ¯ ) < 0 , then u ¯ is a nontrivial solution of (1). Let us take (56) u ¯ = δ sin π t T , where δ will be chosen later. By direct computation, it follows (57) I ( u ¯ ; T ) = t = 1 T [ 1 2 ( ( Δ 3 u ¯ t - 1 ) 2 - A ( Δ 2 u ¯ t - 1 ) 2 + B ( Δ u ¯ t - 1 ) 2 - u ¯ t 2 ) sssssssssss + 1 4 u ¯ t 4 ] = 1 4 δ 2 T ( ( 2 sin π 2 T ) 6 - A ( 2 sin π 2 T ) 4 ssssssssssssss + B ( 2 sin π 2 T ) 2 - 1 + 3 8 δ 2 ) . Since A > 0 if (58) ( 2 sin π 2 T ) 6 + B ( 2 sin π 2 T ) 2 - 1 < 0 for sufficiently small δ , we have I ( u ¯ ) < 0 . We show that there exists T 2 , when T > T 2 , such that (58) is true. Denote ( 2 sin ( π / 2 T ) ) 2 = p , inequality (58) is equivalent to (59) p 3 + B p - 1 < 0 . The roots of the polynomial P 3 ( p ) = p 3 + B p - 1 are (60) p 1 = 1 2 + 1 4 - B 3 27 3 + 1 2 - 1 4 - B 3 27 3 , p 2 = ω 1 2 + 1 4 - B 3 27 3 + ω ¯ 1 2 - 1 4 - B 3 27 3 , p 3 = ω ¯ 1 2 + 1 4 - B 3 27 3 + ω 1 2 - 1 4 - B 3 27 3 , where ω = ( - 1 + 3 i ) / 2 and ω ¯ = ( - 1 - 3 i ) / 2 . Therefore, (59) holds for p < p 1 and (58) holds for 2 sin ( π / 2 T ) < p 1 1 / 2 ; that is, sin ( π / 2 T ) < ( 1 / 2 ) p 1 1 / 2 . Then (1) has a nontrivial solution for T > T 2 , where T 2 is defined as (8).

We can prove the multiplicity result in Theorem 3 using Lemma 9.

Denote B δ the open ball in E T with radius δ about 0; S δ is the boundary of B δ .

Proof of Theorem <xref ref-type="statement" rid="thm1.3">3</xref>.

For integer 1 m T , let us consider the subset K E T , (61) K = { λ 1 sin π t T + λ 2 sin 2 π t T + + λ m sin m π t T : λ 1 2 + λ 2 2 + + λ m 2 = ρ 2 λ 1 sin π t T } , where ρ is an appropriate positive number. Give an odd continuous mapping H : K S ρ : (62) H ( λ 1 sin π t T + λ 2 sin 2 π t T + + λ m sin m π t T ) = ( - λ 1 ρ , - λ 2 ρ , , - λ m ρ ) . As defined in (26), by the monotonicity and superinvariance of genus and Lemma 7, for any ρ > 0 , we have (63) γ ( K S ρ ) = m . K is also a subset of the m -dimensional space (64) E m =    span { sin π t T , sin 2 π t T , , sin m π t T } equipped with the norm (65) λ 1 sin π t T + λ 2 sin 2 π t T + + λ m sin m π t T m 2 = λ 1 2 + λ 2 2 + λ m 2 . Then for every v E m there exist positive constants k 1 and k 2 such that (66) k 1 v m v l 4 k 2 v m . In particular, (67) v l 4 k 2 ρ , v K S ρ . Since T > m T 2 > k T 2 for k = 1,2 , , m , we have (68) 0 < k T m T < 1 T 2 , (69) ( 2 sin k π 2 T ) 6 + B ( 2 sin k π 2 T ) 2 - 1 ( 2 sin m π 2 T ) 6 + B ( 2 sin m π 2 T ) 2 - 1 < ( 2 sin π 2 T 2 ) 6 + B ( 2 sin π 2 T 2 ) 2 - 1 = 0 . For v K S ρ , v t = i = 1 m λ i sin ( i π t / T ) . Using (67) and (69), there holds (70) I ( v ; T ) = t = 1 T [ 1 2 ( ( Δ 3 v t - 1 ) 2 - A ( Δ 2 v t - 1 ) 2 + B ( Δ v t - 1 ) 2 - v t 2 ) sssssssss + 1 4 v t 4 ] t = 1 T [ 1 2 ( ( Δ 3 v t - 1 ) 2 + B ( Δ v t - 1 ) 2 - v t 2 ) + 1 4 v t 4 ] = t = 1 T    [ 1 2 ( ( i = 1 m λ i ( 2 sin i π 2 T ) 3 cos i π ( 2 t + 1 ) 2 T ) 2 sssssssssssslsss + B ( i = 1 m λ i ( 2 sin i π 2 T ) cos i π ( 2 t - 1 ) 2 T ) 2 ssssssssssslssss - ( i = 1 m λ i sin i π t T ) 2 ) + 1 4 v t 4 ] = T 4 ( i = 1 m λ i 2 ( 2 sin i π 2 T ) 6 + B i = 1 m λ i 2 ( 2 sin i π 2 T ) 2 - 1 ) + t = 1 T 1 4 v t 4 ρ 2 4 [ T ( ( 2 sin m π 2 T ) 6 + B ( 2 sin m π 2 T ) 2 - 1 ) ] + k 2 4 ρ 2 < ρ 2 4 [ T ( ( 2 sin π 2 T 2 ) 6 + B ( 2 sin π 2 T 2 ) 2 - 1 ) ] + k 2 4 ρ 2 < 0 for sufficiently small ρ . It follows (71) c 1 < 0 , where c 1 is defined in Lemma 9.

According to Lemma 9, I ( u ; T ) has at least m geometrically distinct critical points. Furthermore, T > m T 2 > T 2 ; similar to the proof of Theorem 2, we draw a conclusion that all m distinct critical points we have obtained are all nontrivial. And the proof of Theorem 3 is completed.

Finally, we exhibit a simple example to illustrate our conclusions.

Example. Consider system (1) with A = 10 and B = 3 .

Solution. Here A = 10 > 0 , B = 3 > 0 , A 2 < 4 B , and (72) sin π 2 T 2 = 1 2 ( 1 2 + 1 4 - B 3 27 3 + 1 2 - 1 4 - B 3 27 3 ) 1 / 2 = 1 2 and it follows we can choose T 2 = 3 / 2 . When T > 3 / 2 , (1) has a nontrivial solution. If one let T > ( 3 / 2 ) m , where 1 m is a integer, then (1) has m distinct nontrivial solution.

Remark. From the given example, one can find our results only depend on coefficients A and B which are very easy to verify and rather relaxed.

Conflict of Interests

The author declares that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

The author thanks the anonymous referee for his/her valuable suggestions. This work was supported by National Natural Science Foundation of China (no. 11101098), the Foundation of Guangzhou Education Bureau (no. 2012A019 ), and PCSIR (no. IRT1226).

Lakshmikantham V. Trigiante D. Theory of Difference Equations: Numerical Methods and Applications 1988 Boston, Mass, USA Academic Press x+242 MR939611 Agarwal R. P. Difference Equations and Inequalities: Theory, Methods, and Applications 1992 New York, NY, USA Marcel Dekker xiv+777 MR1155840 Cai X. C. Yu J. S. Existence of periodic solutions for a 2 n th-order nonlinear difference equation Journal of Mathematical Analysis and Applications 2007 329 2 870 878 10.1016/j.jmaa.2006.07.022 MR2296892 ZBL1153.39302 Guo Z. Yu J. Periodic and subharmonic solutions for superquadratic discrete Hamiltonian systems Nonlinear Analysis. Theory, Methods & Applications A 2003 55 7-8 969 983 10.1016/j.na.2003.07.019 MR2017238 ZBL1053.39011 Deng X. Liu X. Zhang Y. Shi H. Periodic and subharmonic solutions for a 2 n th-order difference equation involving p -Laplacian Indagationes Mathematicae 2013 24 3 613 625 10.1016/j.indag.2013.04.003 MR3064565 Bates P. W. Fife P. C. Gardner R. A. Jones C. K. R. T. The existence of travelling wave solutions of a generalized phase-field model SIAM Journal on Mathematical Analysis 1997 28 1 60 93 10.1137/S0036141095283820 MR1427728 ZBL0867.35041 Caginalp G. Fife P. Higher-order phase field models and detailed anisotropy Physical Review B 1986 34 7 4940 4943 10.1103/PhysRevB.34.4940 MR875362 Peletier L. A. Vorst W. C. Van der Vorst Stationary solutions of a fourthorder nonlinear diffusion equation Differential Equations 1995 31 301 314 Mickens R. E. Difference Equations: Theory and Application 1990 New York, NY, USA Van Nostrand Reinhold xii+448 MR1158461 Kelley W. G. Peterson A. C. Difference Equations: An Introduction with Applications 1991 Boston, Mass, USA Academic Press xii+455 MR1142573 Sharkovsky A. N. Maĭstrenko Yu. L. Romanenko E. Yu. Difference Equations and Their Applications 1993 Dordrecht, The Netherlands Kluwer Academic Publishers xii+358 10.1007/978-94-011-1763-0 MR1244579 Kocić V. L. Ladas G. Global Behavior of Nonlinear Difference Equations of Higher Order with Application 1993 Boston, Mass, USA Kluwer Academic Publishers xii+228 MR1247956 Elaydi S. N. An Introduction to Difference Equations 1996 New York, NY, USA Springer xiv+389 MR1410259 Zhou Z. Yu J. Guo Z. The existence of periodic and subharmonic solutions to subquadratic discrete Hamiltonian systems The ANZIAM Journal 2005 47 1 89 102 10.1017/S1446181100009792 MR2159854 ZBL1081.39019 Yu J. Bin H. Guo Z. Periodic solutions for discrete convex Hamiltonian systems via Clarke duality Discrete and Continuous Dynamical Systems A 2006 15 3 939 950 10.3934/dcds.2006.15.939 MR2220757 ZBL1121.39012 Ahlbrandt C. D. Peterson A. The ( n , n ) -disconjugacy of a 2 n th order linear difference equation Computers & Mathematics with Applications 1994 28 1–3 1 9 10.1016/0898-1221(94)00088-3 MR1284215 Peil T. Peterson A. Asymptotic behavior of solutions of a two term difference equation The Rocky Mountain Journal of Mathematics 1994 24 1 233 252 10.1216/rmjm/1181072463 MR1270038 ZBL0809.39006 Chang K.-C. Infinite-Dimensional Morse Theory and Multiple Solution Problems 1993 Boston, Mass, USA Birkhäuser x+312 MR1196690 Rabinowitz P. H. Minmax Methods in Critical Point Theory with applications to Differential Equations 1986 65 Providence, RI, USA American mathematical Society viii+100 CBMS Regional Conference Series in Mathematics MR845785 Sun J. X. Nonlinear Functional Analysis and Applications 2008 Acedemic Press