Abstract
In this paper, we obtain the global structure of positive solutions for nonlinear discrete simply supported beam equation
with \(f\in C(\mathbb{T}\times [0,\infty ),[0,\infty ))\) satisfying local linear growth condition and \(f(t,0)=0\) uniformly for \(t\in \mathbb{T}\), where \(\mathbb{T}=\{2,\ldots,T\}\), \(\lambda >0\) is a parameter. The main results are based on the global bifurcation theorem.
Similar content being viewed by others
1 Introduction
Difference equations usually describe the evolution of certain phenomena over the course of time, which often occur in numerous settings and forms, both in mathematics and in its applications to economics, statistics, biology, numerical computing, electrical circuit analysis, and other fields; see [1].
It is well known that the fourth-order two-point boundary value problem
appears in the theory of hinged beams [2, 3], so the existence and multiplicity of positive solutions for (1.1) and its discrete analog have been studied by many authors; see, for example, [3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18]. Especially, Refs. [15,16,17,18] introduce the new analytical and numerical methods for differential equations with boundary value problems.
Let a, b be integer and \([a,b]_{\mathbb{Z}}=\{a,a+1,\ldots,b\}\). In 2002, Zhang et al. [10] and He et al. [11] studied the existence of positive solutions for the following discrete analog:
here Δ is forward difference operator with \(\Delta u(t)=u(t+1)-u(t)\), \(h:[2,T]_{\mathbb{Z}}\to [0,\infty )\) and \(f\in C([0,\infty ),[0,\infty ))\). It has been pointed out in [10, 11] that (1.2) is equivalent to the summing equation of the form
where
and
Notice that two distinct Green’s functions are used in the summing equation (1.3), which makes the construction of cones and the verification of strong positivity of \(A_{0}\) become more complex and difficult.
Therefore, Ma and Xu [12] considered the discrete analog of (1.1) as follows:
and introduced the definition of generalized positive solutions.
Definition 1.1
A function \(u:[0,T+2]_{\mathbb{Z}}\to [0, \infty )\) is called a generalized positive solution of (1.4), if u satisfies (1.4), and \(u(t)\geq 0\) on \([1,T+1]_{\mathbb{Z}}\) and \(u(t)>0\) on \([2,T]_{\mathbb{Z}}\).
They also applied the fixed point theorem in cones to obtain some results on the existence of generalized positive solutions of (1.4); see [12]. Ma and Lu [13] applied the Dancer’s global bifurcation theorem to obtain some new results on the existence and multiplicity of generalized positive solutions of discrete simply supported beam equation (1.4) with \(\lambda =1\).
However, in these papers, they assumed that the nonlinearity \(f\in C([2,T]_{\mathbb{Z}}\times [0,\infty ), [0,\infty ))\) satisfies
Of course, a natural question is what would happen if f only is positive on a subinterval \([a,b]_{\mathbb{Z}}\subset [2,T]_{ \mathbb{Z}}\)? That is, f satisfies
Based on the above reasons, we shall show the global structure of positive solutions of (1.4) under the followings assumptions:
-
(H1)
\(f\in C([2,T]_{\mathbb{Z}}\times [0,\infty ),[0,\infty ))\) with \(f(t,0)=0\) for \(t\in \mathbb{T}\);
-
(H2)
there exist \(\alpha _{1}:[2,T]_{\mathbb{Z}}\to (0,\infty )\), \(\beta _{1}:[a,b]_{\mathbb{Z}}\to (0,\infty )\) such that
$$f(t,u)=\alpha _{1}(t)u(t)+\zeta (t,u), \quad (t,u)\in [2,T]_{\mathbb{Z}}\times [0,\infty ), $$and
$$\lim_{u\to \infty }\frac{f(t,u)}{u}=\beta _{1}(t) \quad \text{for a subinterval } t\in [a,b]_{\mathbb{Z}}\subset [2,T]_{\mathbb{Z}}, $$where \(\lim_{u\to 0}\frac{\zeta (t,u)}{u}=0\) uniformly for \(t\in [2,T]_{\mathbb{Z}}\);
-
(H3)
there exist \(\alpha _{2}:[2,T]_{\mathbb{Z}}\to (0,\infty )\), \(\beta _{2}:[a,b]_{\mathbb{Z}}\to (0,\infty )\) such that
$$f(t,u)=\alpha _{2}(t)u(t)+\xi (t,u), \quad (t,u)\in [2,T]_{\mathbb{Z}}\times [0,\infty ), $$and
$$\lim_{u\to 0}\frac{f(t,u)}{u}=\beta _{2}(t) \quad \text{for a subinterval } t\in [a,b]_{\mathbb{Z}}\subset [2,T]_{\mathbb{Z}}, $$where \(\lim_{u\to \infty }\frac{\xi (t,u)}{u}=0 \) uniformly for \(t\in [2,T]_{\mathbb{Z}}\).
Let \(\lambda _{1}(\alpha _{i})\), \(i=1,2\), be the principal eigenvalue of the linear eigenvalue problem
and \(\mu _{1}(\beta _{i})\), \(i=1,2\) the principal eigenvalue of the linear eigenvalue problem
Let \(E=\{u:[0,T+2]_{\mathbb{Z}}\to \mathbb{R}\mid u(1)=u(T+1)=\Delta ^{2}u(0)=\Delta ^{2}u(T)=0\}\) be a Banach space with the norm \(\Vert u \Vert =\max_{t\in [0,T+2]_{\mathbb{Z}}} \vert u(t) \vert \). Denote by Σ the closure of the set
in \(\mathbb{R}\times E\). Let \(\mathbb{E}=\mathbb{R}\times E\) under the product topology. We add the points \(\{(\lambda ,\infty )\mid \lambda \in \mathbb{R}\}\) to our space \(\mathbb{E}\). Let \(S^{+}\) denote the set of generalized positive functions in \(\mathbb{E}\) and \(S^{-}=-S^{+}\), and \(S= S^{-} \cup S^{+}\). They are disjoint and open in \(\mathbb{E}\). Finally, let \(\nu \in \{+,-\}\) and \(\varPhi ^{\nu }=\mathbb{R}\times S ^{\nu }\) and \(\varPhi =\mathbb{R}\times S\).
Our main results are the following.
Theorem 1.1
Let (H1), (H2) hold. If
then there exists a connected component \(\mathscr{C}^{+} \in \varSigma \) such that
-
(i)
\((\mathscr{C}^{+}\backslash \{(\lambda _{1}(\alpha _{1}),0)\} )\subset \varPhi ^{+}\);
-
(ii)
\(\mathscr{C}^{+}\) meets \((\lambda _{1}(\alpha _{1}),0)\) and \((\mu _{1}(\beta _{1}),\infty )\) in λ-direction;
-
(iii)
\(\operatorname{Proj}_{\mathbb{R}} \mathscr{C}^{+}\supset (\mu _{1}(\beta _{1}),\lambda _{1}(\alpha _{1}))\);
-
(iv)
for any \((\lambda , u^{+})\in \mathscr{C}^{+}\), it follows that \(u^{+}(t)>0\), \(t\in [2,T]_{\mathbb{Z}}\).
Corollary 1.1
Let (H1), (H2) hold. If \(\mu _{1}(\beta _{1})< \lambda <\lambda _{1}(\alpha _{1})\), then the problem (1.4) has at least one generalized positive solution.
Theorem 1.2
Let (H1), (H3) hold. If
then there exists a connected component \(\mathscr{D}^{+} \in \varSigma \) such that
-
(i)
\((\mathscr{D}^{+}\backslash \{(\lambda _{1}(\alpha _{2}),\infty )\} )\subset \varPhi ^{\nu }\);
-
(ii)
\(\mathscr{D}^{+}\) meets \((\lambda _{1}(\alpha _{2}),\infty )\) and \((\mu _{1}(\beta _{2}), 0)\) in λ-direction;
-
(iii)
\(\operatorname{Proj}_{\mathbb{R}} \mathscr{D}^{+}\supset (\mu _{1}(\beta _{2}),\lambda _{1}(\alpha _{2}))\);
-
(iv)
for any \((\lambda , u^{+})\in \mathscr{D}^{+}\), it follows that \(u^{+}(t)>0\), \(t\in [2,T]_{\mathbb{Z}}\).
Corollary 1.2
Let (H1), (H3) hold. If \(\mu _{1}(\beta _{2})< \lambda <\lambda _{1}(\alpha _{2})\), then the problem (1.4) has at least one generalized positive solution.
Remark 1.1
Let \(\alpha _{i}(t)\equiv \hat{\alpha }_{i}\) and \(\beta _{i}(t)\equiv \hat{\beta }_{i}\) (\(i=1,2\)) be constants. It is easy to compute that \(\lambda _{1}(\hat{\alpha }_{i})=\frac{(2-2\cos \frac{ \pi }{T})^{2}}{\hat{\alpha }_{i}}\) and \(\mu _{1}(\hat{\beta }_{i})=\frac{(2-2 \cos \frac{\pi }{b+2-a})^{2}}{\hat{\beta }_{i}}\), see [14]. So the conditions of Corollaries 1.1–1.2 are equivalent to
and
respectively.
Moreover, from the above inequalities one concludes that \(\hat{\beta }_{1}=\infty \), \(\hat{\alpha }_{1}=0\), i.e. f is sublinear growth at zero and superlinear growth at infinity about u; \(\hat{\beta }_{2}=\infty \), \(\hat{\alpha }_{2}=0\), i.e. f is superlinear growth at zero and sublinear growth at infinity about u.
Especially, if \([a,b]_{\mathbb{Z}}=[2,T]_{\mathbb{Z}}\), \(\beta _{1}= \alpha _{2}\), \(\beta _{2}=\alpha _{1}\), that is to say, f is linear growth at zero and infinity about u, then the main results give immediately the classical result; see the result of the case \(n=2\) in Theorem 4.1 of [14].
Remark 1.2
If \(\lambda =1\), then the problem (1.4) is the discrete analog of (1.1). Corollaries 1.1–1.2 give the sharp condition of existence results of positive solutions for the discrete analog of (1.1); see [13].
Remark 1.3
It is worth remarking that the global structure of the positive solution curves is very useful for computing the numerical solution of (1.1), for example, it can be used to estimate the value of u in advance in applying the finite difference method.
2 Preliminaries
Let \(Y:=\{u:[0,T+2]_{\mathbb{Z}}\to \mathbb{R}\}\) be the Banach space with the norm \(\Vert u \Vert =\max_{t\in [0,T+2]_{\mathbb{Z}}} \vert u(t) \vert \). Let E be the Banach space
with the norm \(\Vert u \Vert \).
Define a linear operator \(L: E\to Y\) by
From [13], we can see that (1.4) is equivalent to the summing equation
where
It is not difficult to verify that the Green’s function \(G(t,s)\) satisfies the following properties:
where
Moreover, let \([a,b]_{\mathbb{Z}}\) be a subinterval of \([2,T]_{ \mathbb{Z}}\), then, for any \((t,s)\in [a,b]_{\mathbb{Z}}\times [1,T+1]_{ \mathbb{Z}}\), we have
here
Define the cone P in E by
By a standard argument, it is easy to verify that \(T: P \to P\) is completely continuous.
Lemma 2.1
([13, Lemma 3.2])
Let \(\alpha _{i}:[2,T]_{ \mathbb{Z}}\to (0,\infty )\), \(i=1,2\). Then the principal eigenvalue \(\lambda _{1}(\alpha _{i})\) of the eigenvalue problems \((1.5)_{i}\) is positive and the corresponding eigenfunctions \(\varphi _{i}(t)\) is positive on \([2,T]_{\mathbb{Z}}\).
By the same method with obvious changes, we can see that the problem
is equivalent to the summing equation
where
Moreover, by a similar argument to [13, Lemma 3.2], we can obtain the following.
Lemma 2.2
Let \(\beta _{i}:[a,b]_{\mathbb{Z}}\to (0,\infty )\), \(i=1,2\). Then the principal eigenvalue \(\lambda _{1}(\beta _{i})\), \(i=1,2\), of the eigenvalue problems \((1.6)_{i}\) is positive and the corresponding eigenfunctions \(\psi _{i}(t)\) is positive on \([a,b]_{\mathbb{Z}}\).
3 The proof of the main results
To apply the unilateral global bifurcation results [19,20,21,22], we extend f by an odd function \(g:[2,T]_{\mathbb{Z}}\times \mathbb{R}\to \mathbb{R}\) such that
Now let us consider an auxiliary family of equations
It follows from (H2) that
Note that
Proof of Theorem 1.1
(i) Let us consider
as a bifurcation problem for the trivial solution \(u\equiv 0\). Problem (3.4) can be converted to the summing equation
where \(G(t,s)\) is defined by (2.3).
Furthermore, we note that \(\Vert L^{-1}(\zeta (\cdot ,u(\cdot ))) \Vert = \Vert \sum_{s=2}^{T} G(\cdot ,s)\zeta (s,u(s)) \Vert =o( \Vert u \Vert )\) for u near 0 in E, since
From Lemma 2.1, the algebraic multiplicity of \(\lambda _{1}(\alpha _{1})\) equals 1, the pair \((\lambda _{1}(\alpha ),0)\) is a bifurcation point of problem (3.4). Therefore, according to a revised version of [19, Theorem 6.2.1], there exists a component, denoted by \(\mathscr{C} \subset \varSigma \), emanating from \((\lambda _{1}(\alpha _{1}), 0)\). Moreover, (3.4) enjoys all the structural requirements for applying the unilateral global bifurcation theory of López-Gómez [19, Sects. 6.4–6.5], and thanks to the global alternative of Rabinowitz (see, e.g., [20, Corollary 6.3.2]), either \(\mathscr{C}\) is unbounded in \(\mathbb{E}\), or \((\lambda _{j}(\alpha _{1}), 0)\in \mathscr{C}\) for some \(\lambda _{j}(\alpha _{1})\neq \lambda _{1}(\alpha _{1})\), or contains a point \((\lambda ,u)\in \mathbb{R}\times (E_{0}\backslash \{0\})\), here \(\lambda _{j}(\alpha _{1})\) is another eigenvalue of \((1.5)_{1}\), and \(E=\operatorname{span}\{\varphi _{1}\}\oplus E_{0}\).
Although the unilateral bifurcation results of [20, Theorems 1.27 and 1.40] cannot be applied here, among other things because they are false as originally stated (cf. the counterexample of Dancer [21]), the reflection argument of [20] and a similar argument to Theorem 6.4.3 of [19] can be applied to conclude that \(\mathscr{C}=\mathscr{C}^{+}+ \mathscr{C}^{-}\), where \(\mathscr{C}^{+}\) is the component of positive solutions emanating from \((\lambda _{1}(\alpha _{1}),0)\), because of
Moreover, \(\mathscr{C}^{\nu }\) must be unbounded and, \(\mathscr{C} ^{\nu }\backslash \{(\lambda _{1}(\alpha _{1}),0)\}\subset \varPhi ^{+}\).
(ii) It is clear that any solution of (3.4) of the form \((\lambda ,u)\) yields a solution u of (1.4). We will show \(\mathscr{C}^{+}\subset \mathbb{R}\times E\) meets \((\lambda _{1}(\alpha _{1}),0)\) and \((\mu _{1}(\beta _{1}), \infty )\) in λ-direction. To do this, it is enough to show that \(\operatorname{Proj}_{\mathbb{R}}\mathscr{C}^{+}\supset (\mu _{1}(\beta _{1}), \lambda _{1}(\alpha _{1}))\). Let \((\lambda _{n}, u _{n})\in \mathcal{C}^{+}\) satisfy
We note that \(\lambda _{n}>0\) for all \(n\in \mathbb{N}\) since \((0 ,0)\) is the only solution of (3.4) for \(\lambda =0\) and \(\mathscr{C}^{+}\cap (\{0\}\times E)=\emptyset \).
Now we show that
We divide the proof into two steps.
Step 1 We show that there exists a constant number \(M > 0\) such that
Since \((\lambda _{n},u_{n})\) is the solution of (3.4), it follows that
From (H2), there exists a constant \(\varrho _{1}>0\) such that \(\frac{f(t,u)}{u}\geq \varrho _{1} >0\) uniformly for \(t\in [a,b]_{ \mathbb{Z}}\). From (2.2)–(2.5), we have
Then, for any \(t\in [a,b]_{\mathbb{Z}}\), we get
Therefore
Step 2 We show that \(\operatorname{Proj}_{\mathbb{R}}\mathscr{C} ^{+}\supset (\mu _{1}(\beta _{1}), \lambda _{1}(\alpha _{1}))\).
From Step 1 and (3.5), it follows that
This combining (3.7) and (3.8) implies
Let us consider the problem (3.6) and divide (3.6) by \(\Vert u_{n} \Vert \) and set \(v_{n}=\frac{u_{n}}{ \Vert u_{n} \Vert }\). Since \(v_{n}\) is bounded in E, after taking a subsequence if necessary, we have \(v_{n}\to v\) for some \(v\in E\) with \(\Vert v \Vert =1\). Moreover, let \(y_{n}=v_{n}(t)| _{[a,b]_{ \mathbb{Z}}}\) and \(y_{n}\to y=v(t)| _{[a,b]_{\mathbb{Z}}}\), then there exist constants \(r_{1}, r_{2}\geq 0\), \(r_{3}, r_{4}\leq 0\) such that
From (H2), it follows that
where \(\tilde{\lambda }=\lim_{n\to \infty }\lambda _{n}\), \(K(t,s)\) is defined by (2.6), and
again choosing a subsequence and relabeling if necessary. Thus
We claim that \(y\in \mathscr{C}^{+}\). Suppose on the contrary that \(y\notin \mathscr{C}^{+}\). Since \(y\neq 0\) is a solution of (3.10) and there exists \(c\in [a,b]_{\mathbb{Z}}\) such that \(y(c)y(c+1)\leq 0\), which together with the fact \(y_{n}\in E\) implies that y changes its sign in \([a,b]_{\mathbb{Z}}\). This contradicts the facts that \(y_{n}\to y\) in E and \(y_{n}\in \mathscr{C}^{+}\). Therefore \(y\in \mathscr{C}^{+}\). Moreover, let us consider the problem (3.10) and the problem
Multiplying \(\psi _{1}(t)\) in (3.10) and \(y(t)\) in (3.11), then summing from \(t=a\) to b and subtracting, it follows that \(\Delta ^{2}y(b)\psi _{1}(b)+\Delta ^{2}\psi _{1}(b-1)y(b+1)+\psi _{1}(a)\Delta ^{2}y(a-2)+y(a-1) \Delta ^{2}\psi _{1}(a-1) =(\tilde{\lambda }-\mu _{1}(\beta _{1}))\sum_{t=a}^{b}\beta _{1}(t)y(t)\psi _{1}(t)\). It follows from \(y(t)>0\), \(\Delta ^{2}y(t-1)<0\) on \([2,T]_{\mathbb{Z}}\) and \(\psi _{1}(t) \geq 0\), \(\Delta ^{2}\psi _{1}(t-1)\leq 0\) on \([a-1,b+1]_{ \mathbb{Z}}\) that
That is, \(\tilde{\lambda }<\mu _{1}(\beta _{1})\). Thus, \(\operatorname{Proj} _{\mathbb{R}}\mathscr{C}^{+}\supset (\mu _{1}(\beta _{1}), \lambda _{1}( \alpha _{1}))\).
Hence, the conclusions of (ii)–(iv) are true. □
Let \(\lambda _{1}(\alpha _{2})\) is the principal eigenvalue of \((1.5)_{2}\), then from Lemma 2.1, \(\lambda _{1}(\alpha _{2})\) is isolated, having geometric multiplicity 1. Let \(E_{0}\) be a closed subspace of E such that \(E=\operatorname{span}\{\varphi _{2}\}\oplus E_{0}\), where \(\varphi _{2}\) is defined as Lemma 2.1 and \(\Vert \varphi _{2} \Vert =1\). Let \(B_{r}(0)=\{u\in E\mid \Vert u \Vert < r\}\).
Proof of Theorem 1.2
(i) Let \(\xi \in C([2,T]_{\mathbb{Z}} \times \mathbb{R},\mathbb{R})\) be such that
Note that
Let us consider
as a bifurcation problem from the infinity. Problem (3.13) can be converted to the equivalent equation
where \(G(t,s)\) is defined by (2.3). Furthermore, we note that
for u near ∞ in E, since
By a similar argument to [22] and the structural requirements for applying the unilateral global bifurcation theory of López-Gómez [19, Sects. 6.4–6.5], we can conclude to the following.
Let \(\lambda _{1}(\alpha _{2})\) be the principal eigenvalue of \((1.5)_{2}\), such that
for any \(\varepsilon > 0\) small enough, then Σ possesses two unbounded components \(\mathscr{D}^{+} \) and \(\mathscr{D}^{-}\), which meet \((\lambda _{1}(\alpha _{2}),\infty )\). Moreover, if \(\varLambda _{ \ast }\subset \mathbb{R}\) is an interval such that \(\varLambda _{\ast } \cap \varLambda _{2} = \{\lambda _{1}(\alpha _{2})\}\) and \(\mathscr{M}\) is a neighborhood of \((\lambda _{1}(\alpha _{2}),\infty )\) whose projection on \(\mathbb{R}\) lies in \(\varLambda _{\ast }\) and whose projection on E is bounded away from 0, here \(\varLambda _{2}\) denotes the set of real eigenvalues of \((1.5)_{2}\), then at least one of the following three properties is satisfied by \(\mathscr{D}^{\nu }\) for \(\nu \in \{+,-\}\).
-
(1)
\(\mathscr{D}^{\nu }-\mathscr{M}\) is bounded in \(\mathbb{R}\times E\) in which \(\mathscr{D}^{\nu }-\mathscr{M}\) meets \(\{(\lambda ,0)\mid \lambda \in \mathbb{R}\}\) or
-
(2)
\(\mathscr{D}^{\nu }-\mathscr{M}\) is unbounded,
-
(3)
it contains a point \((\lambda ,v)\in \mathbb{R}\times (E_{0} \backslash \{0\})\).
If (2) occurs and \(\mathscr{D}^{\nu }-\mathscr{M}\) has a bounded projection on \(\mathbb{R}\), then \(\mathscr{D}^{\nu }-\mathscr{M}\) meets \((\mu ,\infty )\), where \(\lambda _{1}(\alpha _{2})\neq \mu \in \varLambda _{2}\).
By applying a similar argument to [19, Sects. 6.4–6.5] and [22], it is easy to verify that \(\mathscr{D}^{\nu }\) must be unbounded and, \(\mathscr{D}^{\nu }\subset \{(\lambda _{1}(\alpha _{2}),\infty )\} \cup \varPhi ^{+}\).
(ii) It is clear that any solution of (3.13) of the form \((\lambda ,u)\) yields a solution u of (1.4). We will show \(\mathscr{D}^{+}\) meets \((\lambda _{1}(\alpha _{2}),\infty )\) and \((\mu _{1}(\beta _{2}), 0)\) in the λ-direction. To do this, we only need to show that \(\operatorname{Proj}_{\mathbb{R}}\mathscr{D}^{+} \supset (\mu _{1}(\beta _{2}), \lambda _{1}(\alpha _{2}))\). Let \((\lambda _{n}, u_{n})\in \mathscr{D} ^{+}\) satisfy \(\Vert u_{n} \Vert \to 0\). Then \(\lambda _{n}>0\) for all \(n\in \mathbb{N}\) since \((0 ,0)\) is the only solution of (3.13) for \(\lambda =0\) and \(\mathscr{D}^{+}\cap (\{0\}\times E)=\emptyset \).
Now we show that
By a similar method to proving Step 1 of Theorem 1.1, there exists a constant \(M_{1}\), such that \(0<\lambda _{n}\leq M_{1}\). We only need to prove that
From \(\Vert u_{n} \Vert \to 0\), let us consider (3.9), from (H3), it follows that
where \(\hat{\lambda }=\lim_{n\to \infty }\lambda _{n}\), \(R(t)\) and \(K(t,s)\) are defined as Theorem 1.1, again choosing a subsequence and relabeling if necessary. Thus
We claim that \(y\in \mathscr{D}^{+}\). Suppose on the contrary that \(y\notin \mathscr{D}^{+}\). \(y\neq 0\) is a solution of (3.14); there exists \(c\in [a,b]_{\mathbb{Z}}\) such that \(y(c)y(c+1)\leq 0\), and this together with the fact \(y_{n}\in E\) implies that y changes its sign in \([a,b]_{\mathbb{Z}}\). This contradicts the facts that \(y_{n}\to y\) in E and \(y_{n}\in \mathscr{D}^{+}\). Therefore \(y\in \mathscr{D} ^{+}\). Moreover, let us consider the problem (3.14) and the problem
Multiplying \(\psi _{2}(t)\) in (3.14) and \(y(t)\) in (3.15), then summing from \(t=a\) to b and subtracting, it follows that \(\Delta ^{2}y(b)\psi _{2}(b)+\Delta ^{2}\psi _{2}(b-1)y(b+1)+\psi _{2}(a)\Delta ^{2}y(a-2)+y(a-1) \Delta ^{2}\psi _{2}(a-1) =(\hat{\lambda }-\mu _{1}(\beta _{2}))\sum_{t=a} ^{b}\beta _{2}(t)y(t)\psi _{2}(t)\). It follows from \(y(t)>0\), \(\Delta ^{2}y(t-1)<0\) on \([2,T]_{\mathbb{Z}}\) and \(\psi _{2}(t)\geq 0\), \(\Delta ^{2} \psi _{2}(t-1)\leq 0\) on \([a-1,b+1]_{\mathbb{Z}}\) that
That is, \(\hat{\lambda }<\mu _{1}(\beta _{2})\).
Therefore, the conclusions of (ii)–(iv) are true. □
From the proofs of Theorems 1.1–1.2, we can directly give the conclusions of Corollaries 1.1–1.2.
Example
Let us consider the following problem:
where
and
Clearly, \(f(t,0)=0\) uniformly in \(t\in [1,T]_{\mathbb{Z}}\). Let \(a=2\), \(b=4\), then by computation \(\hat{\beta }_{1}=6\) and \(\mu _{1}( \hat{\beta }_{1})=\frac{6-4\sqrt{2}}{6}<1\), \(\alpha _{1}(t)=g(t)\) and \(\lambda _{1}(\alpha _{1})>1\). From Corollary 1.1 and Remark 1.2, the problem (3.16) has at least one generalized positive solution.
Let us consider the problem (3.16) with the nonlinearity
Clearly, \(f(t,0)=0\) for uniformly \(t\in [2,9]_{\mathbb{Z}}\). Let \(a=2\), \(b=4\), then by simple computation \(\beta _{2}(t)=12-\sqrt{2}-t\), \(\mu _{1}(\beta _{2})<1\) and \(\hat{\alpha }_{2}=\frac{1}{50}\) and \(\lambda _{1}(\hat{\alpha }_{2})=50(2-2\cos \frac{\pi }{9})^{2}>1\). From Corollary 1.2 and Remark 1.2, the problem (3.16) has at least one generalized positive solution.
4 Conclusions
By using the positive property of Green’s function and the unilateral global bifurcation theorem, we obtain the global structure of positive solutions for a class of nonlinear discrete simply supported beam equation with the nonlinearity satisfying local linear growth conditions. The main results extend the existent results of positive solutions and generalize many related problems in the literature.
References
Kelly, W.G., Peterson, A.C.: Difference Equations, an Introduction with Applications, 2nd edn. Academic Press, New York (2001)
Schröder, J.: Operator Inequalities. Mathematics in Science and Engineering, vol. 47. Academic Press, New York (1980)
Agarwal, R.P.: On fourth-order boundary value problems arising in beam analysis. Differ. Integral Equ. 2, 91–110 (1989)
Cabada, A.: The method of lower and upper solutions for second, third, fourth, and higher order boundary value problems. J. Math. Anal. Appl. 185, 302–320 (1994)
Ma, R., Wang, H.: On the existence of positive solutions of fourth-order ordinary differential equations. Appl. Anal. 59, 225–231 (1995)
Bai, Z., Wang, H.: On positive solutions of some nonlinear fourth order beam equations. J. Math. Anal. Appl. 270, 357–368 (2002)
Rynne, B.P.: Infinitely many solutions of superlinear fourth order boundary value problems. Topol. Methods Nonlinear Anal. 19, 303–312 (2002)
Ma, R.: Nodal solutions of boundary value problems of fourth-order ordinary differential equations. J. Math. Anal. Appl. 319, 424–434 (2006)
Webb, J.R.L., Infante, G., Franco, D.: Positive solutions of nonlinear fourth order boundary value problems with local and nonlocal boundary conditions. Proc. R. Soc. Edinb., Sect. A 148, 427–446 (2008)
Zhang, B., Kong, L., Sun, Y., Deng, X.: Existence of positive solutions for BVPs of fourth-order difference equations. Appl. Math. Comput. 131, 583–591 (2002)
He, Z., Yu, J.: On the existence of positive solutions of fourth-order difference equations. Appl. Math. Comput. 161, 139–148 (2005)
Ma, R., Xu, Y.: Existence of positive solution for nonlinear fourth-order difference equations. Comput. Math. Appl. 59, 3770–3777 (2010)
Ma, R., Lu, Y.: Existence and multiplicity of positive solutions of a nonlinear discrete fourth-order boundary value problem. Abstr. Appl. Anal. 2012, 918082 (2012). https://doi.org/10.1155/2012/918082
Gao, C., Xu, J.: Bifurcation techniques and positive solutions of discrete Lidstone boundary value problems. Appl. Math. Comput. 218(2), 434–444 (2011)
Ruggieri, M., Scapellato, A., Speciale, M.P.: Preface of the symposium “Analytical and Numerical Methods for DE with Boundary Value Problems in Applied Sciences”. In: AIP Conference Proceedings, vol. 1978, 140001 (2018). https://doi.org/10.1063/1.5043781.
Barbagallo, A., Ragusa, M.A., Scapellato, A.: Preface of the “5th Symposium on Variational Inequalities and Equilibrium Problems”. In: AIP Conference Proceedings, vol. 1863, 510001 (2017). https://doi.org/10.1063/1.4992659
Jannelli, A., Ruggieri, M., Speciale, M.P.: Analytical and numerical solutions of fractional type advection diffusion equation. In: American Institute of Physics, vol. 1863, 530005 (2017). https://doi.org/10.1063/1.4992675
Scapellato, A.: New perspectives in the theory of some function spaces and their applications. In: AIP Conference Proceedings, vol. 1978, 140002 (2018). https://doi.org/10.1063/1.5043782
López-Gómez, J.: Spectral Theory and Nonlinear Functional Analysis. Research Notes in Mathematics, vol. 426. Chapman & Hall/CRC, Boca Raton (2001)
Rabinowitz, P.: Some global results for nonlinear eigenvalue problems. J. Funct. Anal. 7, 487–513 (1971)
Dancer, E.N.: On the structure of solutions of non-linear eigenvalue problems. Indiana Univ. Math. J. 23, 1069–1076 (1974)
Rabinowitz, P.H.: On bifurcation from infinity. J. Funct. Anal. 14, 462–475 (1973)
Acknowledgements
The authors would like to thank the referees for their useful suggestions which have significantly improved the paper.
Availability of data and materials
Not applicable.
Funding
This work is supported by Natural Science Foundation of China (No. 11801453, No. 11671322), Gansu provincial National Science Foundation of China (No. 1606RJYA232) and NWNU-LKQN-15-16.
Author information
Authors and Affiliations
Contributions
The authors declare that they carried out all the work in this manuscript and read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare that none of them have any competing interests.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Cite this article
Lu, Y., Ma, R. The continuum branch of positive solutions for discrete simply supported beam equation with local linear growth condition. Bound Value Probl 2018, 192 (2018). https://doi.org/10.1186/s13661-018-1113-9
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s13661-018-1113-9