Solutions for a fractional difference boundary value problem

Wei Dong; Jiafa Xu; Donal O’Regan

doi:10.1186/1687-1847-2013-319

Research
Open Access

Solutions for a fractional difference boundary value problem

Wei Dong¹Email author,
Jiafa Xu² and
Donal O’Regan³

Advances in Difference Equations20132013:319

https://doi.org/10.1186/1687-1847-2013-319

Received: 30 July 2013
Accepted: 16 October 2013
Published: 11 November 2013

Abstract

Using a variational approach and critical point theory, we investigate the existence of solutions for a fractional difference boundary value problem.

MSC:26A33, 35A15, 39A12, 44A55.

Keywords

fractional difference boundary value problem
variational approach
critical point theory
solution

where $ν \in (0, 1)$ , ${}_{t}△_{ν - 1}^{ν}$ and ${}_{T}△_{t}^{ν}$ are, respectively, the left fractional difference and the right fractional difference operators, $t \in {[0, T]}_{N_{0}} : = {0, 1, 2, \dots, T}$ , and $f : R \to R$ is continuous.

Fractional calculus has a long history, and there is renewed interest in the study of both fractional calculus and fractional difference equations. In [1, 2], the authors discussed properties of the generalized falling function, a corresponding power rule for fractional delta-operators and the commutativity of fractional sums. A number of papers have appeared which build the theoretical foundations of discrete fractional calculus (for more details, we refer the reader to [3–8] and the references therein).

Atici and Eloe [3] considered the existence of positive solutions for the following two-point boundary value problem for a nonlinear finite fractional difference equation:

{\begin{matrix} - Δ^{ν} y (t) = f (t + ν - 1, y (t + ν - 1)), t = 1, 2, \dots, b + 1, \\ y (ν - 2) = 0, y (ν + b + 1) = 0 . \end{matrix}

(1.2)

In [4], the authors used the mountain pass theorem, a linking theorem, and Clark’s theorem to establish the existence of multiple solutions for a fractional difference boundary value problem with a parameter. Under some suitable assumptions, they obtained some results which ensure the existence of a precise interval of parameters for which the problem admits multiple solutions. We note that there are many papers in the literature [9–18] which discuss discrete problems via variational and critical point theory.

In [9], Tian and Henderson studied the 2n th order nonlinear difference equation

Δ^{n} (r (t - n) Δ^{n} x (t - n)) + f (t, x (t)) = 0, t \in Z,

(1.3)

and established some existence results for anti-periodic solutions under various assumptions on the nonlinearity. In [10], Ye and Tang considered the second-order discrete Hamiltonian system

Δ^{2} u (t - 1) + d (t) {| u (t) |}^{μ - 2} u (t) + \nabla H (t, u (t)) = 0, \forall t \in Z,

and obtained an existence theorem for a nonzero T-periodic solution.

In the literature on discrete problem via critical point theory, the authors are interested in the existence of at least one solution or infinitely many solutions. The existence of a unique solution is not usually studied. In this paper, using Browder’s theorem, first we present a uniqueness result in Section 3. Then a linking theorem is used to establish existence. Finally, assuming an Ambrosetti-Rabinowitz type condition, we show that problem (1.1) has many solutions if the nonlinearity is odd.

2 Preliminaries

For convenience, throughout this paper, we arrange $\sum_{i = j}^{m} x (i) = 0$ for $m < j$ . We present some definitions and lemmas for discrete fractional operators.

For any integer β, let $N_{β} : = {β, β + 1, β + 2, \dots}$ and $t^{(ν)} : = Γ (t + 1) / Γ (t + 1 - ν)$ , where t and ν are determined by (1.1). We also appeal to the convention that if $t + 1 - ν$ is a pole of the gamma function and $t + 1$ is not a pole, then $t^{(ν)} = 0$ .

Definition 2.1 (see [2, 3])

The ν th fractional sum of f for

ν > 0

is defined by

△_{a}^{- ν} f (t) = \frac{1}{Γ (ν)} \sum_{s = a}^{t - ν} {(t - s - 1)}^{(ν - 1)} f (s) for t \in N_{a - ν} .

(2.1)

We also define the ν th fractional difference for $ν > 0$ by $△^{ν} f (t) : = △^{N} △^{ν - N} f (t)$ , where $t \in N_{a + N - ν}$ and $N \in N$ is chosen so that $0 \leq N - 1 < ν \leq N$ .

Definition 2.2 (see [2, 3])

Let f be any real-valued function and

ν \in (0, 1)

. The left discrete fractional difference and the right discrete fractional difference operators are, respectively, defined as

\begin{aligned} \begin{aligned} {}_{t}△_{a}^{ν} f (t) & = △_{t} △_{a}^{- (1 - ν)} f (t) \\ = \frac{1}{Γ (1 - ν)} △ \sum_{s = a}^{t + ν - 1} {(t - s - 1)}^{(- ν)} f (s), t \equiv a - ν + 1 (mod 1), \end{aligned} \\ \begin{aligned} {}_{b}△_{t}^{ν} f (t) & = - △_{b} △_{t}^{- (1 - ν)} f (t) \\ = \frac{1}{Γ (1 - ν)} (- △) \sum_{s = t + 1 - ν}^{b} {(s - t - 1)}^{(- ν)} f (s), t \equiv b + ν - 1 (mod 1) . \end{aligned} \end{aligned}

(2.2)

Definition 2.3 Suppose that X is a Banach space and

I : X \to R

is a functional defined on X. For given

x, y \in X

, assume that

lim_{ε \to 0} \frac{I (x + ε y) - I (x)}{ε}

(2.3)

exists. Then I is Gateaux differentiable at x, the limit in (2.3) is called the Gateaux differential of I at x in direction y, and is denoted by

(I^{'} (x), y)

, i.e.,

(I^{'} (x), y) = lim_{ε \to 0} \frac{I (x + ε y) - I (x)}{ε} .

(2.4)

Definition 2.4 (see [[19], p.303])

Let X be a reflexive real Banach space and $X^{*}$ its dual. The operator $L : X \to X^{*}$ is said to be demicontinuous if L maps strongly convergent sequences in X to weakly convergent sequences in $X^{*}$ .

Lemma 2.5 (Browder theorem, see [[19], Theorem 5.3.22])

Let X be a reflexive real Banach space. Moreover, let

L : X \to X^{*}

be an operator satisfying the conditions

(i)

L is bounded and demicontinuous,
(ii)

L is coercive, i.e., ${lim}_{∥ x ∥ \to \infty} \frac{(L (x), x)}{∥ x ∥} = + \infty$ ,
(iii)

L is monotone on the space X, i.e., for all $x, y \in X$ , we have
$(L (x) - L (y), x - y) \geq 0 .$

(2.5)

Then the equation $L (x) = f^{*}$ has at least one solution $x \in X$ for every $f^{*} \in X^{*}$ . If, moreover, the inequality (2.5) is strict for all $x, y \in X$ , $x \neq y$ , then the equation $L (x) = f^{*}$ has precisely one solution $x \in X$ for all $f^{*} \in X^{*}$ .

Definition 2.6 (see [20–22])

Let X be a real Banach space, $I \in C^{1} (X, R)$ and $c \in R$ . We say that I satisfies the (PS)_c condition if any sequence ${x_{n}} \subset X$ such that $I (x_{n}) \to c$ and $I^{'} (x_{n}) \to 0$ as $n \to \infty$ has a convergent subsequence.

Lemma 2.7 (Linking theorem, Rabinowitz, see [20–22])

Let

X = Y \oplus Z

be a Banach space with Z closed in X and

dim Y < \infty

. Let

ρ > r > 0

, and let

z \in Z

be such that

∥ z ∥ = r

. Define

\begin{matrix} M : = {u = y + λ z : ∥ u ∥ \leq ρ, λ \geq 0, y \in Y}, N : = {u \in Z : ∥ u ∥ = r}, \\ M_{0} : = {u = y + λ z : y \in Y, ∥ u ∥ = ρ and λ \geq 0, or ∥ u ∥ \leq ρ and λ = 0} . \end{matrix}

Let

I \in C^{1} (X, R)

be such that

b : = inf_{u \in N} I (u) > a : = max_{u \in M_{0}} I (u) .

If I satisfies the (PS)_c condition with

c : = inf_{γ \in Γ} max_{u \in M} I (γ (u)), where Γ : = {γ \in C (M, X) : γ |_{M_{0}} = id},

then c is a critical point of I.

Definition 2.8 (see [20–22])

Let X be a real Banach space and $I \in C^{1} (X, R)$ . We say that I satisfies the Cerami condition ((C) condition for short) if any sequence ${x_{n}} \subset X$ such that $I (x_{n})$ is bounded and $(1 + ∥ x_{n} ∥) ∥ I^{'} (x_{n}) ∥ \to 0$ as $n \to \infty$ , there exists a subsequence of ${x_{n}}$ which is convergent in X.

Lemma 2.9 (Mountain pass theorem, see [20–22])

Let X be a real Banach space, and let

I \in C^{1} (X, R)

satisfy the (C) condition. If

I (θ) = 0

and the following conditions hold:

(i)

there are two positive constants ρ, η and a closed linear subspace $X_{1}$ of X such that codim $X_{1} = l$ and $I |_{X_{1} \cap \partial B_{ρ}} \geq η$ , where $B_{ρ}$ is an open ball of radius ρ with center θ;
(ii)

there is a subspace $X_{2}$ with $dim X_{2} = m$ , $m > l$ , such that
$I (x) \to - \infty as ∥ x ∥ \to \infty, x \in X_{2} .$

Then I possesses at least $m - l$ distinct pairs of nontrivial critical points.

In what follows, we establish the variational framework for (1.1). Let

X : = {x = {(x (ν - 1), x (ν), \dots, x (ν + T - 1))}^{+} : x (ν + i - 1) \in R, i = 0, 1, \dots, T} .

(2.6)

Then X is the

T + 1

-dimensional Hilbert space with the usual inner product and the usual norm

(x, z) = \sum_{t = ν - 1}^{T + ν - 1} x (t) z (t), ∥ x ∥ = {(\sum_{t = ν - 1}^{T + ν - 1} {| x (t) |}^{2})}^{\frac{1}{2}}, x, z \in X .

(2.7)

For

α > 1

, we define the α-norm on X:

{∥ x ∥}_{α} = {(\sum_{t = ν - 1}^{T + ν - 1} {| x (t) |}^{α})}^{\frac{1}{α}}

. Since

dim X < \infty

, we see that there exist

c_{1 α} > 0

,

c_{2 α} > 0

such that

c_{1 α} ∥ x ∥ \leq {∥ x ∥}_{α} \leq c_{2 α} ∥ x ∥

(2.8)

for all x belonging to X (or its subspace).

In view of [[4], (3.4)], we can define an energy functional on X by

I (x) = \frac{1}{2} \sum_{t = - 1}^{T} (_{t} △_{ν - 1}^{ν} x (t)) 2 - \sum_{t = - 1}^{T} F (x (t + ν - 1)), x \in X,

(2.9)

where

\begin{matrix} F (x (t + ν - 1)) = \int_{0}^{x (t + ν - 1)} f (s) d s, \\ x (ν - 2) = 0, [_{t} △_{ν - 1}^{ν} x (t)] t = T = \frac{- ν}{Γ (1 - ν)} \sum_{s = ν - 1}^{T + ν} {(T + s - 1)}^{(- ν - 1)} x (s) = 0 . \end{matrix}

Clearly,

I (θ) = 0

. Let

\begin{aligned} E : = & {χ = {(x (ν - 2), x (ν - 1), \dots, x (ν + T))}^{+} \\ \in R^{T + 3} : x (ν - 2) = 0, [_{t} △_{ν - 1}^{ν} x (t)] t = T = 0} . \end{aligned}

(2.10)

Then, from the boundary conditions of (1.1), it is easy to see that E is isomorphic to X. In the following, when we say

x \in X

, we always imply that x can be extended to

χ \in E

if it is necessary. Now we claim that if

x = {(x (ν - 1), x (ν), \dots, x (ν + T - 1))}^{+} \in X

is a critical point of I, then

χ = {(x (ν - 2), x (ν - 1), \dots, x (ν + T))}^{+} \in E

is precisely a solution of (1.1). Indeed, since I can be viewed as a continuously differentiable functional defined on the finite dimensional Hilbert space X, the Fréchet derivative

I^{'} (x)

is zero if and only if

\partial I (x) / \partial x (i) = 0

for all

i = ν - 1, ν, \dots, ν + T - 1

. From the relation between the Fréchet derivative and the Gateaux derivative, we obtain

\begin{aligned} (I^{'} (x), y) = & lim_{ε \to 0} \frac{I (x + ε y) - I (x)}{ε} = \frac{1}{ε} [\frac{1}{2} \sum_{t = - 1}^{T} {[(_{t} △_{ν - 1}^{ν} (x (t) + ε y (t)))}^{2} - (_{t} △_{ν - 1}^{ν} x (t)) 2] \\ - \sum_{t = - 1}^{T} [F ((x + ε y) (t + ν - 1)) - F (x (t + ν - 1))]] \\ = & {\sum_{t = - 1}^{T}}_{t} △_{ν - 1}^{ν} x {(t)}_{t} △_{ν - 1}^{ν} y (t) - \sum_{t = - 1}^{T} f (x (t + ν - 1)) y (t + ν - 1) . \end{aligned}

(2.11)

Therefore, in order to obtain the existence of solutions for (1.1), we only need to study the existence of critical points of the energy functional I on X.

Next, noting Definition 2.2, for

t \in {[- 1, T]}_{N_{- 1}}

, we let

{}_{t}△_{ν - 1}^{ν} x (t) = △ \frac{1}{Γ (1 - ν)} \sum_{s = ν - 1}^{t - (1 - ν)} {(t - s - 1)}^{(- ν)} x (s) : = △ z (t + ν - 1) .

(2.12)

Then we have

\begin{aligned} z (ν - 2) = 0, \\ z (ν - 1) = \frac{1}{Γ (1 - ν)} \sum_{s = ν - 1}^{0 - (1 - ν)} {(- s - 1)}^{(- ν)} x (s) = x (ν - 1), \\ z (ν) = \frac{1}{Γ (1 - ν)} \sum_{s = ν - 1}^{1 - (1 - ν)} {(1 - s - 1)}^{(- ν)} x (s) = (1 - ν) x (ν - 1) + x (ν), \\ z (ν + 1) = \frac{1}{Γ (1 - ν)} \sum_{s = ν - 1}^{2 - (1 - ν)} {(2 - s - 1)}^{(- ν)} x (s) \\ = \frac{(2 - ν) (1 - ν)}{2!} x (ν - 1) + (1 - ν) x (ν) + x (ν + 1), \\ ⋮ \\ z (ν + T - 1) = \frac{1}{Γ (1 - ν)} \sum_{s = ν - 1}^{T - (1 - ν)} {(T - s - 1)}^{(- ν)} x (s) \\ = \frac{(T - ν) (T - 1 - ν) \dots (1 - ν)}{T!} x (ν - 1) \\ + \frac{(T - 1 - ν) (T - 2 - ν) \dots (1 - ν)}{(T - 1)!} x (ν) + \dots \\ + (1 - ν) x (ν + T - 2) + x (ν + T - 1), \end{aligned}

i.e.,

z = B x

, where

z = {(z (ν - 1), z (ν), \dots, z (ν + T - 1))}^{+}

,

x = {(x (ν - 1), x (ν), \dots, x (ν + T - 1))}^{+}

,

B = {(\begin{array}{ccccc} 1 & 0 & 0 & \dots & 0 \\ 1 - ν & 1 & 0 & \dots & 0 \\ \frac{(2 - ν) (1 - ν)}{2!} & 1 - ν & 1 & \dots & 0 \\ ⋮ & ⋮ & ⋮ & ⋮ & ⋮ \\ \frac{(T - ν) (T - 1 - ν) \dots (1 - ν)}{T!} & \frac{(T - 1 - ν) (T - 2 - ν) \dots (1 - ν)}{(T - 1)!} & \dots & \dots & 1 \end{array})}_{(T + 1) \times (T + 1)} .

Clearly,

{(B^{- 1})}^{+} B^{- 1}

is a positive definite matrix. All the eigenvalues of

{(B^{- 1})}^{+} B^{- 1}

are positive. Let

λ_{min}

and

λ_{max}

denote respectively the minimum and the maximum eigenvalues of

{(B^{- 1})}^{+} B^{- 1}

. Since

x = B^{- 1} z

, we have

λ_{min} {∥ z ∥}^{2} \leq {∥ x ∥}^{2} = (z^{+} {(B^{- 1})}^{+}, B^{- 1} z) \leq λ_{max} {∥ z ∥}^{2} .

(2.13)

Next, let

A = {(\begin{array}{cccccc} 2 & - 1 & 0 & \dots & 0 & 0 \\ - 1 & 2 & - 1 & \dots & 0 & 0 \\ 0 & - 1 & 2 & \dots & 0 & 0 \\ ⋮ & ⋮ & ⋮ & ⋮ & ⋮ & ⋮ \\ 0 & 0 & 0 & \dots & 2 & - 1 \\ 0 & 0 & 0 & \dots & - 1 & 1 \end{array})}_{(T + 1) \times (T + 1)} .

By direct verification, we see that A is a positive definite matrix. Let $η_{1}, \dots, η_{T + 1}$ be the orthonormal eigenvectors corresponding to the eigenvalues $λ_{1}, \dots, λ_{T + 1}$ of A, where $0 < λ_{1} < \dots < λ_{T + 1}$ . Clearly, $X : = span {η_{1}, \dots, η_{T + 1}}$ . Let $Y : = span {η_{1}, \dots, η_{l}}$ , $Z : = span {η_{l + 1}, \dots, η_{T + 1}}$ , for $l \in {[1, T + 1]}_{N_{0}}$ . Then $X = Y \oplus Z$ .

3 Main results

Now we state our main results and give their proof. For convenience, we list assumptions on f and F:

(H1)

$f (θ) \neq 0$ , and there is a constant c such that $| f (x) - f (y) | \leq c | x - y |$ , $\forall x, y \in X$ .
(H2)

There exist a constant $d > 0$ and $μ \in (1, 2)$ such that ${lim sup}_{| x | \to 0} \frac{F (x)}{{| x |}^{μ}} < d$ .
(H3)

$\frac{λ_{l}}{2 λ_{min}} x^{2} \leq \int_{0}^{x} f (s) d s$ for all $x \in R$ .
(H4)

There exist $α > 2$ and $R > 0$ such that
$0 < α \int_{0}^{x} f (s) d s \leq x f (x) for | x | > R .$
(H5)

There is a constant $α > 2$ such that ${lim inf}_{| x | \to \infty} \frac{F (x)}{{| x |}^{α}} > 0$ .
(H6)

There is $γ > 1$ such that
$\underset{| x | \to \infty}{lim inf} \frac{f (x) x - 2 F (x)}{{| x |}^{γ}} > 0 .$
(H7)

$f (- x) + f (x) = 0$ , $\forall x \in R$ .

Theorem 3.1 Let (H1) hold. Then (1.1) has precisely one solution for $c \in (0, \frac{λ_{1}}{λ_{max}})$ .

Proof We shall apply Lemma 2.5 to prove the result. From (2.11), we define the operator

(L (x), y) = {\sum_{t = - 1}^{T}}_{t} △_{ν - 1}^{ν} x {(t)}_{t} △_{ν - 1}^{ν} y (t) - \sum_{t = - 1}^{T} f (x (t + ν - 1)) y (t + ν - 1), \forall x, y \in X .

(3.1)

Clearly, if for all

y \in X

, there exists

x_{0} \in X

such that

(L (x_{0}), y) = 0

, then

x_{0}

is a solution of (1.1). Let

\begin{aligned} (L_{1} (x), y) = {\sum_{t = - 1}^{T}}_{t} △_{ν - 1}^{ν} x {(t)}_{t} △_{ν - 1}^{ν} y (t), \\ (L_{2} (x), y) = \sum_{t = - 1}^{T} f (x (t + ν - 1)) y (t + ν - 1), \forall x, y \in X . \end{aligned}

We sketch the properties of

L_{1}

and

L_{2}

. It is clear that

L_{1}

is a linear operator, and furthermore,

L_{1}

is bounded. Indeed, the Cauchy-Schwarz inequality enables us to obtain, notice (2.12) and (2.13),

\begin{aligned} | (L_{1} (x), y) | & \leq \sum_{t = - 1}^{T} |_{t} △_{ν - 1}^{ν} x (t) {| |}_{t} △_{ν - 1}^{ν} y (t) | \\ \leq {(\sum_{t = - 1}^{T} |_{t} △_{ν - 1}^{ν} x (t) |^{2})}^{\frac{1}{2}} {(\sum_{t = - 1}^{T} |_{t} △_{ν - 1}^{ν} y (t) |^{2})}^{\frac{1}{2}} \\ \leq \frac{λ_{T + 1}}{λ_{min}} ∥ x ∥ ∥ y ∥ < \infty, x, y \in X . \end{aligned}

(3.2)

Consequently,

L_{1}

is continuous on X. Next, we show that

L_{2}

is bounded and continuous. Let

y = θ

in (H1) and

| f (θ) | = c_{1} > 0

. Then we have from (H1)

| f (x) | \leq c | x | + c_{1}, \forall x \in X .

(3.3)

From (3.3), the definition of

L_{2}

, and the Cauchy-Schwarz inequality, we obtain

\begin{aligned} | (L_{2} (x), y) | & \leq \sum_{t = 0}^{T} | f (x (t + ν - 1)) | | y (t + ν - 1) | \\ \leq \sum_{t = 0}^{T} (c | x (t + ν - 1) | + c_{1}) | y (t + ν - 1) | \\ \leq c ∥ x ∥ ∥ y ∥ + c_{1} \sqrt{T + 1} ∥ y ∥ < \infty, \forall x, y \in X, \end{aligned}

(3.4)

\begin{aligned} | (L_{2} (x_{1}) - L_{2} (x_{2}), y) | & \leq \sum_{t = 0}^{T} | f (x_{1} (t + ν - 1)) - f (x_{2} (t + ν - 1)) | | y (t + ν - 1) | \\ \leq \sum_{t = 0}^{T} c | x_{1} (t + ν - 1) - x_{2} (t + ν - 1) | | y (t + ν - 1) | \\ \leq c ∥ x_{1} - x_{2} ∥ ∥ y ∥, \forall x, y \in X . \end{aligned}

(3.5)

Therefore, $L_{2}$ is bounded and continuous, as required. Hence, L is bounded and continuous, so demicontinuous.

From (3.3), notice (2.12) and (2.13), we see

\begin{aligned} (L (x), x) & = \sum_{t = - 1}^{T} (_{t} △_{ν - 1}^{ν} x (t)) 2 - \sum_{t = - 1}^{T} f (x (t + ν - 1)) x (t + ν - 1) \\ = \sum_{t = - 1}^{T - 1} {(△ z (t + ν - 1))}^{2} - \sum_{t = 0}^{T} f (x (t + ν - 1)) x (t + ν - 1) \\ \geq \frac{λ_{1}}{λ_{max}} {∥ x ∥}^{2} - \sum_{t = 0}^{T} (c | x (t + ν - 1) | + c_{1}) | x (t + ν - 1) | \\ \geq (\frac{λ_{1}}{λ_{max}} - c) {∥ x ∥}^{2} - c_{1} \sqrt{T + 1} ∥ x ∥ . \end{aligned}

(3.6)

Therefore, ${lim}_{∥ x ∥ \to \infty} \frac{(L (x), x)}{∥ x ∥} = + \infty$ , i.e., L is coercive on X.

Finally, we prove that L is strictly monotone. Indeed, from (H1), we have

\begin{aligned} (L (x) - L (y), x - y) \geq & \sum_{t = - 1}^{T} (_{t} △_{ν - 1}^{ν} x (t) -_{t} △_{ν - 1}^{ν} y (t)) 2 \\ - \sum_{t = 0}^{T} | f (x (t + ν - 1)) - f (y (t + ν - 1)) | | x (t + ν - 1) - y (t + ν - 1) | \\ \geq & \frac{λ_{1}}{λ_{max}} {∥ x - y ∥}^{2} - c {∥ x - y ∥}^{2} > 0, for x, y \in X and x \neq y . \end{aligned}

(3.7)

All the conditions of Lemma 2.5 are satisfied, as claimed. Hence, (1.1) has precisely one solution. This completes the proof. □

Theorem 3.2 Let (H2)-(H4) hold. Then (1.1) has at least one solution.

Proof From (H2), there exists

δ > 0

with

F (x) \leq d {| x |}^{μ} for | x | \leq δ .

(3.8)

Thus, for

x \in Z

with

∥ x ∥ \leq δ

, it follows from the Hölder inequality that

\begin{aligned} I (x) & = \frac{1}{2} \sum_{t = - 1}^{T} (_{t} △_{ν - 1}^{ν} x (t)) 2 - \sum_{t = - 1}^{T} F (x (t + ν - 1)) \\ = \frac{1}{2} \sum_{t = - 1}^{T - 1} {(Δ z (t + ν - 1))}^{2} - \sum_{t = 0}^{T} F (x (t + ν - 1)) \\ \geq \frac{λ_{l + 1}}{2} {∥ z ∥}^{2} - \sum_{t = 0}^{T} d {| x (t + ν - 1) |}^{μ} \\ \geq \frac{λ_{l + 1}}{2 λ_{max}} {∥ x ∥}^{2} - d {(\sum_{t = 0}^{T} {| x (t + ν - 1) |}^{2})}^{\frac{μ}{2}} {(\sum_{t = 0}^{T} 1)}^{1 - \frac{μ}{2}} \\ = \frac{λ_{l + 1}}{2 λ_{max}} {∥ x ∥}^{2} - d {(T + 1)}^{1 - \frac{μ}{2}} {∥ x ∥}^{μ} . \end{aligned}

(3.9)

By virtue of the inequality

1 < μ < 2

, there exists

r > 0

such that

b : = inf_{∥ x ∥ = r, x \in Z} I (x) > 0 .

(3.10)

From (H3), for

x \in Y

, we see

\begin{aligned} I (x) & = \frac{1}{2} \sum_{t = - 1}^{T} (_{t} △_{ν - 1}^{ν} x (t)) 2 - \sum_{t = - 1}^{T} F (x (t + ν - 1)) \\ = \frac{1}{2} \sum_{t = - 1}^{T - 1} {(Δ z (t + ν - 1))}^{2} - \sum_{t = 0}^{T} F (x (t + ν - 1)) \\ \leq \frac{λ_{l}}{2 λ_{min}} {∥ x ∥}^{2} - \sum_{t = 0}^{T} F (x (t + ν - 1)) \\ = \sum_{t = 0}^{T} [\frac{λ_{l}}{2 λ_{min}} x^{2} (t + ν - 1) - F (x (t + ν - 1))] \leq 0 . \end{aligned}

(3.11)

From (H4), we see that there exist

c_{2}, c_{3} > 0

such that

\int_{0}^{x} f (s) d s \geq c_{2} {| x |}^{α} - c_{3}, \forall x \in R .

(3.12)

Hence, for

x \in X

, we find

\begin{aligned} I (x) & = \frac{1}{2} \sum_{t = - 1}^{T} (_{t} △_{ν - 1}^{ν} x (t)) 2 - \sum_{t = - 1}^{T} F (x (t + ν - 1)) \\ = \frac{1}{2} \sum_{t = - 1}^{T - 1} {(Δ z (t + ν - 1))}^{2} - \sum_{t = 0}^{T} F (x (t + ν - 1)) \\ \leq \frac{λ_{T + 1}}{2 λ_{min}} {∥ x ∥}^{2} - c_{2} \sum_{t = 0}^{T} {| x (t + ν - 1) |}^{α} + c_{3} (T + 1) . \end{aligned}

(3.13)

Set

z : = r \frac{η_{l + 1}}{∥ η_{l + 1} ∥}

with

r > 0

is given in (3.10). For

Y \oplus R_{z} \subset X

, (2.8) holds true. This, together with (3.13), implies

I (x) \leq \frac{λ_{T + 1}}{2 λ_{min}} {∥ x ∥}^{2} - c_{2} c_{1 α}^{α} {∥ x ∥}^{α} + c_{3} (T + 1) .

Since

α > 2

, we obtain

lim_{∥ x ∥ \to \infty, x \in Y \oplus R_{z}} I (x) = - \infty .

(3.14)

Define

\begin{matrix} M : = {x = y + λ z : ∥ x ∥ \leq ρ, λ \geq 0, y \in Y}, \\ M_{0} : = {x = y + λ z : y \in Y, ∥ x ∥ = ρ and λ \geq 0, or ∥ x ∥ \leq ρ and λ = 0} . \end{matrix}

Since

z \in Z

and then

I (z) \geq b > 0

, (3.11) and (3.14) guarantee that there is

ρ > r

such that

a : = max_{x \in M_{0}} I (x) \leq 0 .

It remains to prove that I satisfies the (PS)_c condition. This will be the case if we show that any sequence

{x_{n}}_{n = 1}^{\infty} \subset X

such that

d : = sup_{n} I (x_{n}) < \infty, I^{'} (x_{n}) \to 0,

contains a convergent subsequence. Note that

dim X < \infty

, so we only need to show the boundedness of

{x_{n}}_{n = 1}^{\infty}

. Take

β > 0

such that

β^{- 1} \in (2, α)

for n large enough, and (H4), (3.12) and (2.8) enable us to obtain

\begin{aligned} d + ∥ x_{n} ∥ \geq & I (x_{n}) - \frac{1}{β} (I^{'} (x_{n}), x_{n}) \\ = & (\frac{1}{2} - \frac{1}{β}) \sum_{t = - 1}^{T} (_{t} △_{ν - 1}^{ν} x_{n} (t)) 2 \\ + \sum_{t = - 1}^{T} (\frac{1}{β} f (x_{n} (t + ν - 1)) x_{n} (t + ν - 1) - F (x_{n} (t + ν - 1))) \\ \geq & (\frac{1}{2} - \frac{1}{β}) \frac{λ_{1}}{λ_{max}} {∥ x_{n} ∥}^{2} + \sum_{t = 0}^{T} (\frac{α}{β} - 1) F (x_{n} (t + ν - 1)) \\ \geq & (\frac{1}{2} - \frac{1}{β}) \frac{λ_{1}}{λ_{max}} {∥ x_{n} ∥}^{2} + \sum_{t = 0}^{T} (\frac{α}{β} - 1) (c_{2} {| x_{n} (t + ν - 1) |}^{α} - c_{3}) \\ \geq & (\frac{1}{2} - \frac{1}{β}) \frac{λ_{1}}{λ_{max}} {∥ x_{n} ∥}^{2} + (\frac{α}{β} - 1) c_{2} c_{1 α}^{α} {∥ x_{n} ∥}^{α} - (\frac{α}{β} - 1) (T + 1) c_{3} . \end{aligned}

Since $α > 2$ and $(\frac{α}{β} - 1) > 0$ , we see that ${x_{n}}_{n = 1}^{\infty}$ is bounded.

Thus the functional I satisfies all the conditions of Lemma 2.7, and then I has a critical point, and (1.1) has at least one solution. This completes the proof. □

Theorem 3.3 Let (H2), (H5)-(H7) hold. Then (1.1) has at least $m - l$ solutions.

Proof We shall utilize Lemma 2.9 to prove the result. If

X_{1} = Z = span {η_{l + 1}, \dots, η_{T + 1}}

, we see codim

X_{1} = l

. From (H2), noting (3.9), we can take

ρ = {(d d_{1} λ_{max} λ_{l + 1}^{- 1} {(T + 1)}^{1 - \frac{μ}{2}})}^{\frac{1}{2 - μ}}

so that

ρ \leq δ

, where

d_{1} > 2

. Therefore,

\begin{aligned} I_{X_{1} \cap \partial B_{ρ}} & \geq {(d d_{1} λ_{max} λ_{l + 1}^{- 1} {(T + 1)}^{1 - \frac{μ}{2}})}^{\frac{μ}{2 - μ}} [\frac{λ_{l + 1}}{2 λ_{max}} d d_{1} λ_{max} λ_{l + 1}^{- 1} {(T + 1)}^{1 - \frac{μ}{2}} - d {(T + 1)}^{1 - \frac{μ}{2}}] \\ = (\frac{d_{1}}{2} - 1) d {(d d_{1} λ_{max} λ_{l + 1}^{- 1} {(T + 1)}^{1 - \frac{μ}{2}})}^{\frac{μ}{2 - μ}} {(T + 1)}^{1 - \frac{μ}{2}} . \end{aligned}

Thus (i) of Lemma 2.9 holds true.

Choose

X_{2} : = span {η_{1}, \dots, η_{m}}

, where

m > l

, and

dim X_{2} = m

. From (H5), we see that there exist

c_{4}, c_{5} > 0

such that

F (x) \geq c_{4} {| x |}^{α} - c_{5}, \forall x \in R .

Therefore, from (3.13) and (2.8), we arrive at

\begin{aligned} - I (x) & \geq - \frac{λ_{m}}{2 λ_{min}} {∥ x ∥}^{2} + c_{4} \sum_{t = 0}^{T} {| x (t + ν - 1) |}^{α} - c_{5} (T + 1) \\ \geq - \frac{λ_{m}}{2 λ_{min}} {∥ x ∥}^{2} + c_{4} c_{1 α}^{α} {∥ x ∥}^{α} - c_{5} (T + 1) . \end{aligned}

Since $α > 2$ , $I (x) \to - \infty$ as $∥ x ∥ \to \infty$ , $x \in X_{2}$ . Thus (ii) of Lemma 2.9 holds true.

Finally, we prove that I satisfies the (C) condition. Let

{x_{n}} \subset X

be such that for some

M_{1} > 0

,

| I (x_{n}) | \leq M_{1}, (1 + ∥ x_{n} ∥) ∥ I^{'} (x_{n}) ∥ \to 0 as n \to \infty .

We claim that

∥ x_{n} ∥

is bounded. Otherwise, suppose that

∥ x_{n} ∥ \to \infty

as

n \to \infty

. It is easy to see that for any

n \in N

, there exists

M_{2}

such that

2 I (x_{n}) - (I^{'} (x_{n}), x_{n}) \leq M_{2} .

On the other hand, from (H6), there exist

c_{6}, c_{7} > 0

such that

f (x) x - 2 F (x) \geq c_{6} {| x |}^{γ} - c_{7}, \forall x \in R .

Consequently, from (2.8),

\begin{aligned} 2 I (x_{n}) - (I^{'} (x_{n}), x_{n}) & = \sum_{t = 0}^{T} [f (x_{n} (t + ν - 1)) x_{n} (t + ν - 1) - 2 F (x_{n} (t + ν - 1))] \\ \geq \sum_{t = 0}^{T} (c_{6} {| x_{n} (t + ν - 1) |}^{γ} - c_{7}) \geq c_{6} c_{1 γ}^{γ} {∥ x_{n} ∥}^{γ} - c_{7} (T + 1) . \end{aligned}

Let $n \to \infty$ , and we get a contradiction.

It is easy to see that I is even and $I (θ) = 0$ . Thus all the conditions of Lemma 2.9 are satisfied, and (1.1) has at least $m - l$ solutions. The proof is complete. □

Examples

1.

Let $f (x) = η_{1} x + η_{2}$ , where $η_{1} \in (0, \frac{λ_{1}}{λ_{max}})$ and $η_{2} \neq 0$ . Clearly, (H1) holds.
2.

Let $f (x) = \frac{λ_{l}}{λ_{min}} (x + x^{3}) e^{x^{2}} + 4 x^{3}$ . Then $F (x) = \frac{λ_{l}}{2 λ_{min}} x^{2} e^{x^{2}} + x^{4}$ . Thus, (H2) and (H3) hold automatically. For $R \geq 1$ , $α = 4$ , we see
$0 < \frac{2 λ_{l}}{λ_{min}} x^{2} e^{x^{2}} + 4 x^{4} \leq x f (x) = \frac{λ_{l}}{λ_{min}} (x^{2} + x^{4}) e^{x^{2}} + 4 x^{4} for all | x | > 1 .$

Therefore, (H4) holds.
3.

Let $f (x) = 2 x + 4 x^{3}$ . Then $F (x) = x^{2} + x^{4}$ and (H2), (H7) hold. Choose $α = 4$ , $γ = 4$ , and we see
$\underset{| x | \to \infty}{lim inf} \frac{x^{2} + x^{4}}{{| x |}^{4}} = 1 > 0, \underset{| x | \to \infty}{lim inf} \frac{f (x) x - 2 F (x)}{{| x |}^{4}} = \underset{| x | \to \infty}{lim inf} \frac{2 x^{4}}{{| x |}^{4}} = 2 > 0 .$

Hence (H5) and (H6) hold.

Declarations

Acknowledgements

The authors sincerely thank the reviewers for their valuable suggestions and useful comments that have led to the present improved version of the original manuscript. Research is supported by the NNSF-China (10971046 and 11371117), Shandong and Hebei Provincial Natural Science Foundation (ZR2012AQ007 and A2012402036), GIIFSDU (yzc12063) and IIFSDU (2012TS020).

Authors’ Affiliations

(1)

Department of Mathematics, Hebei University of Engineering, Handan, Hebei, 056038, China

(2)

School of Mathematics, Shandong University, Jinan, Shandong, 250100, China

(3)

School of Mathematics, Statistics and Applied Mathematics, National University of Ireland, Galway, Ireland

References

Atici FM, Eloe PW: A transform method in discrete fractional calculus. Int. J. Differ. Equ. 2007, 2: 165-176.MathSciNetGoogle Scholar
Atici FM, Şengül S: Modeling with fractional difference equations. J. Math. Anal. Appl. 2010, 369: 1-9. 10.1016/j.jmaa.2010.02.009MathSciNetView ArticleGoogle Scholar
Atici FM, Eloe PW: Two point boundary value problems for finite fractional difference equations. J. Differ. Equ. Appl. 2011, 17: 445-456. 10.1080/10236190903029241MathSciNetView ArticleGoogle Scholar
Xie ZS, Jin YF, Hou CM: Multiple solutions for a fractional difference boundary value problem via variational approach. Abstr. Appl. Anal. 2012., 2012: Article ID 143914Google Scholar
Goodrich CS: On discrete sequential fractional boundary value problems. J. Math. Anal. Appl. 2012, 385: 111-124. 10.1016/j.jmaa.2011.06.022MathSciNetView ArticleGoogle Scholar
Goodrich CS: On a fractional boundary value problem with fractional boundary conditions. Appl. Math. Lett. 2012, 25: 1101-1105. 10.1016/j.aml.2011.11.028MathSciNetView ArticleGoogle Scholar
Lv WD: Existence of solutions for discrete fractional boundary value problems with a p -Laplacian operator. Adv. Differ. Equ. 2012., 2012: Article ID 163Google Scholar
Pan YY, Han ZL, Sun SR, Huang ZQ: The existence and uniqueness of solutions to boundary value problems of fractional difference equations. Math. Sci. 2012., 2012: Article ID 6Google Scholar
Tian Y, Henderson J: Anti-periodic solutions of higher order nonlinear difference equations: a variational approach. J. Differ. Equ. Appl. 2013, 19: 1380-1392. 10.1080/10236198.2012.752467MathSciNetView ArticleGoogle Scholar
Ye YW, Tang CL: Periodic solutions for second-order discrete Hamiltonian system with a change of sign in potential. Appl. Math. Comput. 2013, 219: 6548-6555. 10.1016/j.amc.2013.01.019MathSciNetView ArticleGoogle Scholar
Deng XQ, Liu X, Zhang YB, Shi HP: Periodic and subharmonic solutions for a 2 n th-order difference equation involving p -Laplacian. Indag. Math. 2013, 24: 613-625. 10.1016/j.indag.2013.04.003MathSciNetView ArticleGoogle Scholar
Zhang GD, Sun HR: Multiple solutions for a fourth-order difference boundary value problem with parameter via variational approach. Appl. Math. Model. 2012, 36: 4385-4392. 10.1016/j.apm.2011.11.064MathSciNetView ArticleGoogle Scholar
Mawhin J: Periodic solutions of second order nonlinear difference systems with ϕ -Laplacian: a variational approach. Nonlinear Anal. 2012, 75: 4672-4687. 10.1016/j.na.2011.11.018MathSciNetView ArticleGoogle Scholar
Zhang X, Shi YM: Homoclinic orbits of a class of second-order difference equations. J. Math. Anal. Appl. 2012, 396: 810-828. 10.1016/j.jmaa.2012.07.016MathSciNetView ArticleGoogle Scholar
Wang SL, Liu JS: Nontrivial solutions of a second order difference systems with multiple resonance. Appl. Math. Comput. 2012, 218: 9342-9352. 10.1016/j.amc.2012.03.017MathSciNetView ArticleGoogle Scholar
Iannizzotto A, Tersian SA: Multiple homoclinic solutions for the discrete p -Laplacian via critical point theory. J. Math. Anal. Appl. 2013, 403: 173-182. 10.1016/j.jmaa.2013.02.011MathSciNetView ArticleGoogle Scholar
Bereanu C, Jebelean P, Şerban C:Periodic and Neumann problems for discrete $p (\cdot)$ -Laplacian. J. Math. Anal. Appl. 2013, 399: 75-87. 10.1016/j.jmaa.2012.09.047MathSciNetView ArticleGoogle Scholar
Galewski M, Smejda J: On the dependence on parameters for mountain pass solutions of second order discrete BVP’s. Appl. Math. Comput. 2013, 219: 5963-5971. 10.1016/j.amc.2012.12.028MathSciNetView ArticleGoogle Scholar
Drábek P, Milota J: Methods of Nonlinear Analysis: Applications to Differential Equations. Birkhäuser, Basel; 2007.Google Scholar
Willem M: Minimax Theorems. Birkhäuser, Boston; 1996.View ArticleGoogle Scholar
Struwe M: Variational Methods: Applications to Nonlinear Partial Differential Equations and Hamiltonian Systems. Springer, Berlin; 2008.Google Scholar
Rabinowitz PH: Minimax Methods in Critical Point Theory with Applications to Differential Equations. Am. Math. Soc., Providence; 1986.Google Scholar

Copyright

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/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.