Theory and Modern Applications

# Existence of solutions for nonlinear second-order q-difference equations with first-order q-derivatives

## Abstract

In this paper, we establish the existence of solutions for a boundary value problem with the nonlinear second-order q-difference equation

$\left\{\begin{array}{l}{D}_{q}^{2}u\left(t\right)=f\left(t,u\left(t\right),{D}_{q}u\left(t\right)\right),\phantom{\rule{1em}{0ex}}t\in I,\\ {D}_{q}u\left(0\right)=0,\phantom{\rule{2em}{0ex}}{D}_{q}u\left(1\right)=\alpha u\left(1\right).\end{array}$

The existence and uniqueness of solutions for the problem are proved by means of the Leray-Schauder nonlinear alternative and some standard fixed point theorems. Finally, we give two examples to demonstrate the use of the main results. The nonlinear team f contains ${D}_{q}u\left(t\right)$ in the equation.

## 1 Introduction

In this paper, we study the existence of solutions for a boundary value problem with nonlinear second-order q-difference equations

$\left\{\begin{array}{l}{D}_{q}^{2}u\left(t\right)=f\left(t,u\left(t\right),{D}_{q}u\left(t\right)\right),\phantom{\rule{1em}{0ex}}t\in I,\\ {D}_{q}u\left(0\right)=0,\phantom{\rule{2em}{0ex}}{D}_{q}u\left(1\right)=\alpha u\left(1\right),\end{array}$
(1.1)

where $f\in C\left(I×{R}^{2},R\right)$, $I=\left\{{q}^{n}:n\in N\right\}\cup \left\{0,1\right\}$, $q\in \left(0,1\right)$, and $\alpha \ne 0$ is a fixed real number.

The q-difference equations initiated at the beginning of the twentieth century  is a very interesting field in difference equations. In the last few decades, it has evolved into a multidisciplinary subject and plays an important role in several fields of physics such as cosmic strings and black holes , conformal quantum mechanics , nuclear and high energy physics . However, the theory of boundary value problems (BVPs) for nonlinear q-difference equations is still in the initial stages and many aspects of this theory need to be explored. To the best of our knowledge, for the BVPs of nonlinear q-difference equations, a few works were done; see  and the references therein. In particular, the study of BVPs for nonlinear q-difference equation with first-order q-difference is yet to be initiated.

The main aim of this paper is to develop some existence and uniqueness results for BVP (1.1). Our results are based on a variety of fixed point theorems such as the Banach contraction mapping principle, the Leray-Schauder nonlinear alternative and the Leray-Schauder continuous theorem. Some examples and special cases are also discussed.

## 2 Preliminary results

In this section, firstly, let us recall some basic concepts of q-calculus [14, 15].

Definition 2.1 For $0, we define the q-derivative of a real-value function f as

${D}_{q}f\left(t\right)=\frac{f\left(t\right)-f\left(qt\right)}{\left(1-q\right)t},\phantom{\rule{2em}{0ex}}{D}_{q}f\left(0\right)=\underset{t\to 0}{lim}{D}_{q}f\left(t\right).$

Note that ${lim}_{q\to {1}^{-}}{D}_{q}f\left(t\right)={f}^{\prime }\left(t\right)$.

Definition 2.2 The higher-order q-derivatives are defined inductively as

${D}_{q}^{0}f\left(t\right)=f\left(t\right),\phantom{\rule{2em}{0ex}}{D}_{q}^{n}f\left(t\right)={D}_{q}{D}_{q}^{n-1}f\left(t\right),\phantom{\rule{1em}{0ex}}n\in N.$

For example, ${D}_{q}\left({t}^{k}\right)={\left[k\right]}_{q}{t}^{k-1}$, where k is a positive integer and the bracket ${\left[k\right]}_{q}=\left({q}^{k}-1\right)/\left(q-1\right)$. In particular, ${D}_{q}\left({t}^{2}\right)=\left(1+q\right)t$.

Definition 2.3 The q-integral of a function f defined in the interval $\left[a,b\right]$ is given by

${\int }_{a}^{x}f\left(t\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}t:=\sum _{n=0}^{\mathrm{\infty }}x\left(1-q\right){q}^{n}f\left(x{q}^{n}\right)-af\left({q}^{n}a\right),\phantom{\rule{1em}{0ex}}x\in \left[a,b\right],$

and for $a=0$, we denote

${I}_{q}f\left(x\right)={\int }_{0}^{x}f\left(t\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}t=\sum _{n=0}^{\mathrm{\infty }}x\left(1-q\right){q}^{n}f\left(x{q}^{n}\right).$

Then

${\int }_{a}^{b}f\left(t\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}t={\int }_{0}^{b}f\left(t\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}t-{\int }_{0}^{a}f\left(t\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}t.$

Similarly, we have

${I}_{q}^{0}f\left(t\right)=f\left(t\right),\phantom{\rule{2em}{0ex}}{I}_{q}^{n}f\left(t\right)={I}_{q}{I}_{q}^{n-1}f\left(t\right),\phantom{\rule{1em}{0ex}}n\in N.$

Observe that

${D}_{q}{I}_{q}f\left(x\right)=f\left(x\right),$

and if f is continuous at $x=0$, then ${I}_{q}{D}_{q}f\left(x\right)=f\left(x\right)-f\left(0\right)$.

In q-calculus, the product rule and integration by parts formula are

${D}_{q}\left(gh\right)\left(t\right)={D}_{q}g\left(t\right)h\left(t\right)+g\left(qt\right){D}_{q}h\left(t\right),$
(2.1)
${\int }_{0}^{x}f\left(t\right){D}_{q}g\left(t\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}t={\left[f\left(t\right)g\left(t\right)\right]}_{0}^{x}-{\int }_{0}^{x}{D}_{q}f\left(t\right)g\left(qt\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}t.$
(2.2)

Remark 2.4 In the limit $q\to {1}^{-}$, the above results correspond to their counterparts in standard calculus.

Definition 2.5 $f:I×{R}^{2}\to R$ is called an S-Carathéodory function if and only if

1. (i)

for each $\left(u,v\right)\in {R}^{2}$, $t↦f\left(t,u,v\right)$ is measurable on I;

2. (ii)

for a.e. $t\in I$, $\left(u,v\right)↦f\left(t,u,v\right)$ is continuous on ${R}^{2}$;

3. (iii)

for each $r>0$, there exists ${\phi }_{r}\left(t\right)\in {L}^{1}\left(I,{R}^{+}\right)$ with $t{\phi }_{r}\left(t\right)\in {L}^{1}\left(I,{R}^{+}\right)$ on I such that $max\left\{|u|,|v|\right\}\le r$ implies $|f\left(t,u,v\right)|\le {\phi }_{r}\left(t\right)$, for a.e. I, where , and normed by ${\parallel u\parallel }_{{L}^{1}}={\int }_{0}^{1}|u\left(t\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}t$ for all $u\in {L}^{1}\left(I,{R}^{+}\right)$.

Theorem 2.6 (Nonlinear alternative for single-valued maps )

Let E be a Banach space, let C be a closed and convex subset of E, and let U be an open subset of C and $0\in U$. Suppose that $F:\overline{U}\to C$ is a continuous, compact (that is, $F\left(\overline{U}\right)$ is a relatively compact subset of C) map. Then either

1. (i)

F has a fixed point in $\overline{U}$, or

2. (ii)

there is a $u\in \partial U$ (the boundary of U in C) and $\lambda \in \left(0,1\right)$ with $u=\lambda F\left(u\right)$.

Lemma 2.7 Let $y\in C\left[0,1\right]$, then the BVP

$\left\{\begin{array}{l}{D}_{q}^{2}u\left(t\right)=y\left(t\right),\phantom{\rule{1em}{0ex}}t\in I,\\ {D}_{q}u\left(0\right)=0,\phantom{\rule{2em}{0ex}}{D}_{q}u\left(1\right)=\alpha u\left(1\right),\end{array}$
(2.3)

has a unique solution

$\begin{array}{rl}u\left(t\right)& ={\int }_{0}^{t}\left(t-qs\right)y\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s+{\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)y\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ ={\int }_{0}^{1}G\left(t,s;q\right)y\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s,\end{array}$
(2.4)

where

$G\left(t,s;q\right)=\frac{1}{\alpha }\left\{\begin{array}{ll}1-\alpha +\alpha t,& s\le t,\\ 1-\alpha +\alpha qs,& t\le s.\end{array}$
(2.5)

Proof Integrating the q-difference equation from 0 to t, we get

${D}_{q}u\left(t\right)={\int }_{0}^{t}y\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s+{a}_{1}.$
(2.6)

Integrating (2.6) from 0 to t and changing the order of integration, we have

$u\left(t\right)={\int }_{0}^{t}\left(t-qs\right)y\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s+{a}_{1}t+{a}_{0},$
(2.7)

where ${a}_{1}$, ${a}_{0}$ are arbitrary constants. Using the boundary conditions ${D}_{q}u\left(0\right)=0,{D}_{q}u\left(1\right)=\alpha u\left(1\right)$ in (2.7), we find that ${a}_{1}=0$, and

${a}_{1}={\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)y\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s.$

Substituting the values of ${a}_{0}$ and ${a}_{1}$ in (2.7), we obtain

$u\left(t\right)={\int }_{0}^{t}\left(t-qs\right)y\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s+{\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)y\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s.$

This completes the proof. □

Remark 2.8 For $q\to 1$, equation (2.4) takes the form

$u\left(t\right)={\int }_{0}^{t}\left(t-qs\right)y\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s+{a}_{1}t+{a}_{0},$

which is the solution of a classical second-order ordinary differential equation ${u}^{″}\left(t\right)=y\left(t\right)$ and the associated form of Green’s function for the classical case is

$G\left(t,s\right)=\frac{1}{\alpha }\left\{\begin{array}{ll}1-\alpha +\alpha t,& s\le t,\\ 1-\alpha +\alpha s,& t\le s.\end{array}$

We consider the Banach space ${C}_{q}=C\left(I,R\right)$ equipped with the standard norm $\parallel u\parallel =max\left\{{\parallel u\parallel }_{\mathrm{\infty }},{\parallel {D}_{q}u\parallel }_{\mathrm{\infty }}\right\}$, and ${\parallel \cdot \parallel }_{\mathrm{\infty }}=sup\left\{\parallel \cdot \parallel ,t\in I\right\}$, $u\in {C}_{q}$.

Define an integral operator $T:{C}_{p}\to {C}_{p}$ by

$\begin{array}{rl}Tu\left(t\right)=& {\int }_{0}^{1}G\left(t,s;q\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ =& {\int }_{0}^{t}\left(t-qs\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ +{\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s,\phantom{\rule{1em}{0ex}}t\in I,u\in {C}_{q}.\end{array}$
(2.8)

Obviously, T is well defined and $u\in {C}_{q}$ is a solution of BVP (1.1) if and only if u is a fixed point of T.

## 3 Existence and uniqueness results

In this section, we apply various fixed point theorems to BVP (1.1). First, we give the uniqueness result based on Banach’s contraction principle.

Theorem 3.1 Let $f:I×{R}^{2}\to R$ be a continuous function, and there exists ${L}_{1}\left(t\right),{L}_{2}\left(t\right)\in C\left(\left[0,1\right],\left[0,+\mathrm{\infty }\right)\right)$ such that

$|f\left(t,{u}_{1},{v}_{1}\right)-f\left(t,{u}_{2},{v}_{2}\right)|\le {L}_{1}\left(t\right)|{u}_{1}-{u}_{2}|+{L}_{2}\left(t\right)|{v}_{1}-{v}_{2}|,\phantom{\rule{1em}{0ex}}t\in I,\left({u}_{1},{v}_{1}\right),\left({u}_{2},{v}_{2}\right)\in {R}^{2}.$

(${\mathrm{H}}_{\mathrm{1}}$) $\mathrm{\Lambda }<|\alpha |$ for $0<|\alpha |<1$, or

(${\mathrm{H}}_{\mathrm{2}}$) $\mathrm{\Lambda }<1$ for $|\alpha |\ge 1$

holds, where $\mathrm{\Lambda }={max}_{t\in \left[0,1\right]}\left\{{L}_{1}\left(t\right)+{L}_{2}\left(t\right)\right\}$. Then BVP (1.1) has a unique solution.

Proof Case 1: $|\alpha |<1$. Let us set ${sup}_{t\in I}|f\left(t,0,0\right)|={M}_{0}$ and choose

$r\ge \frac{{M}_{0}}{|\alpha |\left(1-\delta \right)},$
(3.1)

where δ is such that $\frac{\mathrm{\Lambda }}{|\alpha |}\le \delta \le 1$. Now we show that $T{B}_{r}\subset {B}_{r}$, where ${B}_{r}=\left\{u\in {C}_{q}:\parallel u\parallel \le r\right\}$. For each $u\in {B}_{r}$, we have

$\begin{array}{rl}|Tu\left(t\right)|\le & \underset{t\in I}{sup}|{\int }_{0}^{t}\left(t-qs\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s+{\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le & \underset{t\in I}{sup}|{\int }_{0}^{t}\left(t-qs\right)\left(|f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)-f\left(s,0,0\right)|+|f\left(s,0,0\right)|\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ +{\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)\left(|f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)-f\left(s,0,0\right)|+|f\left(s,0,0\right)|\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le & \underset{t\in I}{sup}|{\int }_{0}^{t}\left(t-qs\right)\left({L}_{1}\left(s\right)|u\left(s\right)|+{L}_{2}\left(s\right)|{D}_{q}u\left(s\right)|+|f\left(s,0,0\right)|\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ +{\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)\left({L}_{1}\left(s\right)|u\left(s\right)|+{L}_{2}\left(s\right)|{D}_{q}u\left(s\right)|+|f\left(s,0,0\right)|\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le & \left(\mathrm{\Lambda }\parallel u\parallel +{M}_{0}\right)\underset{t\in I}{sup}|{\int }_{0}^{t}\left(t-qs\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s+{\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le & \left(\mathrm{\Lambda }\parallel u\parallel +{M}_{0}\right)\underset{t\in I}{sup}\left\{|\frac{{t}^{2}}{1+q}+\frac{1}{\alpha }-1+\frac{q}{1+q}|\right\}\\ \le & \left(\mathrm{\Lambda }\parallel u\parallel +{M}_{0}\right)\frac{1}{|\alpha |}\le \left(\mathrm{\Lambda }r+{M}_{0}\right)\frac{1}{|\alpha |}\le \left(\frac{\mathrm{\Lambda }}{|\alpha |}+\left(1-\delta \right)\right)r\le r,\end{array}$

and

$\begin{array}{rl}|{D}_{q}Tu\left(t\right)|& \le |{D}_{q}Tu\left(t\right)|\le \underset{t\in I}{sup}|{\int }_{0}^{t}f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le \underset{t\in I}{sup}{\int }_{0}^{t}\left(|f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)-f\left(s,0,0\right)|+|f\left(s,0,0\right)|\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ \le \underset{t\in I}{sup}{\int }_{0}^{t}\left({L}_{1}\left(s\right)|u\left(s\right)|+{L}_{2}\left(s\right)|{D}_{q}u\left(s\right)|+|f\left(s,0,0\right)|\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ \le \left(\mathrm{\Lambda }\parallel u\parallel +{M}_{0}\right)\le \left(\mathrm{\Lambda }r+|\alpha |\left(1-\delta \right)r\right)\le \left(\frac{\mathrm{\Lambda }}{|\alpha |}+\left(1-\delta \right)\right)r\le r.\end{array}$

Hence, we obtain that $\parallel Tu\parallel \le r$, so $T{B}_{r}\subset {B}_{r}$.

Now, for $u,v\in {C}_{q}$ and for each $t\in I$, we have

$\begin{array}{rl}|Tu\left(t\right)-Tv\left(t\right)|\le & \underset{t\in I}{sup}|Tu\left(t\right)-Tv\left(t\right)|\\ \le & \underset{t\in I}{sup}|{\int }_{0}^{t}\left(t-qs\right)|f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)-f\left(s,v\left(s\right),{D}_{q}v\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ +{\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)|f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)-f\left(s,v\left(s\right),{D}_{q}v\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le & \underset{t\in I}{sup}|{\int }_{0}^{t}\left(t-qs\right)\left({L}_{1}\left(s\right)|u\left(s\right)-v\left(s\right)|+{L}_{2}\left(s\right)|{D}_{q}u\left(s\right)-{D}_{q}v\left(s\right)|\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ +{\int }_{0}^{1}\left(\frac{1}{\alpha }-1+qs\right)\left({L}_{1}\left(s\right)|u\left(s\right)-v\left(s\right)|+{L}_{2}\left(s\right)|{D}_{q}u\left(s\right)-{D}_{q}v\left(s\right)|\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le & \mathrm{\Lambda }\underset{t\in I}{sup}\left\{|\frac{{t}^{2}}{1+q}+\frac{1}{\alpha }-1+\frac{q}{1+q}|\right\}\parallel u-v\parallel \\ \le & \frac{\mathrm{\Lambda }}{|\alpha |}\parallel u-v\parallel <\parallel u-v\parallel ,\end{array}$

and

$\begin{array}{rl}|{D}_{q}Tu\left(t\right)-{D}_{q}Tv\left(t\right)|& \le \underset{t\in I}{sup}|{D}_{q}Tu\left(t\right)-{D}_{q}Tv\left(t\right)|\\ \le \underset{t\in I}{sup}|{\int }_{0}^{t}|f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)-f\left(s,v\left(s\right),{D}_{q}v\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le \underset{t\in I}{sup}|{\int }_{0}^{t}\left({L}_{1}\left(s\right)|u\left(s\right)-v\left(s\right)|+{L}_{2}\left(s\right)|{D}_{q}u\left(s\right)-{D}_{q}v\left(s\right)|\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le \mathrm{\Lambda }\parallel u-v\parallel \le \frac{\mathrm{\Lambda }}{|\alpha |}\parallel u-v\parallel <\parallel u-v\parallel .\end{array}$

Therefore, we obtain that $\parallel Tu-Tv\parallel <\parallel u-v\parallel$, so T is a contraction. Thus, the conclusion of the theorem follows by Banach’s contraction mapping principle.

Case 2: $|\alpha |\ge 1$. It is similar to the proof of case 1. This completes the proof of Theorem 3.1. □

Corollary 3.2 Assume that $f:I×{R}^{2}\to R$ is a continuous function and there exist two positive constants ${L}_{1}$, ${L}_{2}$ such that

$|f\left(t,{u}_{1},{v}_{1}\right)-f\left(t,{u}_{2},{v}_{2}\right)|\le {L}_{1}|{u}_{1}-{u}_{2}|+{L}_{2}|{v}_{1}-{v}_{2}|,\phantom{\rule{1em}{0ex}}t\in I,\left({u}_{1},{v}_{1}\right),\left({u}_{2},{v}_{2}\right)\in {R}^{2}.$

(${\mathrm{H}}_{\mathrm{3}}$) ${L}_{1}+{L}_{2}<|\alpha |$ for $0<|\alpha |<1$, or

(${\mathrm{H}}_{\mathrm{4}}$) ${L}_{1}+{L}_{2}<1$ for $|\alpha |\ge 1$

holds. Then BVP (1.1) has a unique solution.

Corollary 3.3 Assume that $f:I×{R}^{2}\to R$ is a continuous function and there exist two functions ${L}_{1}\left(t\right),{L}_{2}\left(t\right)\in {L}^{1}\left(I,{R}^{+}\right)$ such that

$|f\left(t,{u}_{1},{v}_{1}\right)-f\left(t,{u}_{2},{v}_{2}\right)|\le {L}_{1}|{u}_{1}-{u}_{2}|+{L}_{2}|{v}_{1}-{v}_{2}|,\phantom{\rule{1em}{0ex}}t\in I,\left({u}_{1},{v}_{1}\right),\left({u}_{2},{v}_{2}\right)\in {R}^{2}.$

(${\mathrm{H}}_{\mathrm{5}}$) $A+Bq|\alpha |<|\alpha |$ for $0<|\alpha |<1$, or

(${\mathrm{H}}_{\mathrm{6}}$) $A<1$ for $|\alpha |\ge 1$

holds, where

$A={\int }_{0}^{1}\left[{L}_{1}\left(s\right)+{L}_{2}\left(s\right)\right]\phantom{\rule{0.2em}{0ex}}{d}_{q}s,\phantom{\rule{2em}{0ex}}B={\int }_{0}^{1}s\left[{L}_{1}\left(s\right)+{L}_{2}\left(s\right)\right]\phantom{\rule{0.2em}{0ex}}{d}_{q}s.$

Then BVP (1.1) has a unique solution.

Proof It is similar to the proof of Theorem 3.1. □

The next existence result is based on the Leray-Schauder nonlinear alternative theorem.

Lemma 3.4 Let $f:I×{R}^{2}\to R$ be an S-Carathéodory function. Then $T:{C}_{q}\to {C}_{q}$ is completely continuous.

Proof The proof consists of several steps.

1. (i)

T maps bounded sets into bounded sets in ${C}_{q}$.

Let ${B}_{r}=\left\{u\in {C}_{q}:\parallel u\parallel \le r\right\}$ be a bounded set in ${C}_{q}$ and $u\in {B}_{r}$. Then we have

$\begin{array}{rl}|Tu\left(t\right)|& \le {\int }_{0}^{t}|t-qs||f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s+{\int }_{0}^{1}|\frac{1}{\alpha }-1+qs||f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ \le \left(1+\frac{1}{|\alpha |}\right){\int }_{0}^{1}{\phi }_{r}\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s=\left(1+\frac{1}{|\alpha |}\right){\parallel {\phi }_{r}\parallel }_{{L}^{1}},\end{array}$

and

$|{D}_{q}Tu\left(t\right)|\le {\int }_{0}^{t}|f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s\le {\int }_{0}^{1}{\phi }_{r}\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s={\parallel {\phi }_{r}\parallel }_{{L}^{1}}.$

Thus $\parallel Tu\parallel \le max\left\{{\parallel Tu\parallel }_{\mathrm{\infty }},{\parallel {D}_{q}Tu\parallel }_{\mathrm{\infty }}\right\}\le \left(1+\frac{1}{|\alpha |}\right){\parallel {\phi }_{r}\parallel }_{{L}^{1}}$.

1. (ii)

T maps bounded sets into equicontinuous sets of ${C}_{q}$.

Let ${r}_{1},{r}_{2}\in I$, ${r}_{1}<{r}_{2}$, and let ${B}_{r}$ be a bounded set of ${C}_{q}$ as before. Then, for $u\in {B}_{r}$, we have

$\begin{array}{rl}|Tu\left({r}_{2}\right)-Tu\left({r}_{1}\right)|=& |{\int }_{0}^{{r}_{2}}\left({r}_{2}-qs\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ -{\int }_{0}^{{r}_{1}}\left({r}_{1}-qs\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ =& |{\int }_{0}^{{r}_{1}}\left({r}_{2}-{r}_{1}\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ +{\int }_{{r}_{1}}^{{r}_{2}}\left({r}_{2}-qs\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le & {\int }_{0}^{{r}_{1}}|{r}_{2}-{r}_{1}|{\phi }_{r}\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s+{\int }_{{r}_{1}}^{{r}_{2}}|{r}_{2}-qs|{\phi }_{r}\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\to 0\\ \left({r}_{2}-{r}_{1}\to 0\right),\end{array}$

and

$\begin{array}{rl}|\phantom{\rule{0.2em}{0ex}}{d}_{q}Tu\left({r}_{2}\right)-{D}_{q}Tu\left({r}_{1}\right)|=& |{\int }_{0}^{{r}_{2}}f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s-{\int }_{0}^{{r}_{1}}f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ =& |{\int }_{{r}_{1}}^{{r}_{2}}f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\le {\int }_{{r}_{1}}^{{r}_{2}}{\phi }_{r}\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\to 0\\ \left({r}_{2}-{r}_{1}\to 0\right).\end{array}$

As a consequence of the Arzelá-Ascoli theorem, we can conclude that $T:{C}_{q}\to {C}_{q}$ is completely continuous. This proof is completed. □

Theorem 3.5 Let $f:I×{R}^{2}\to R$ be an S-Carathéodory function. Suppose further that there exists a real number $M>0$ such that

$\frac{|\alpha |M}{\left(1+|\alpha |\right){\parallel {\phi }_{r}\parallel }_{{L}^{1}}}>1$

holds, where

${\parallel {\phi }_{r}\parallel }_{{L}^{1}}={\int }_{0}^{1}{\phi }_{r}\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\ne 0.$

Then BVP (1.1) has at least one solution.

Proof In view of Lemma 3.4, we obtain that $T:{C}_{q}\to {C}_{q}$ is completely continuous. Let $\lambda \in \left(0,1\right)$ and $u=\lambda Tu$. Then, for $t\in I$, we have

$\begin{array}{rl}|u\left(t\right)|=& |\lambda Tu\left(t\right)|\\ \le & {\int }_{0}^{t}|t-qs||f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ +{\int }_{0}^{1}|\frac{1}{\alpha }-1+qs||f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ \le & \left(1+\frac{1}{|\alpha |}\right){\int }_{0}^{1}{\phi }_{r}\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s=\left(1+\frac{1}{|\alpha |}\right){\parallel {\phi }_{r}\parallel }_{{L}^{1}},\end{array}$

and

$|{D}_{q}u\left(t\right)|=|{D}_{q}\lambda Tu\left(t\right)|\le {\int }_{0}^{t}|f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s\le {\int }_{0}^{1}{\phi }_{r}\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s={\parallel {\phi }_{r}\parallel }_{{L}^{1}}.$

Hence, consequently,

$\frac{|\alpha |\parallel u\parallel }{\left(1+|\alpha |\right){\parallel {\phi }_{r}\parallel }_{{L}^{1}}}\le 1.$

Therefore, there exists $M>0$ such that $\parallel u\parallel \ne M$. Let us set $U=\left\{u\in {C}_{q}:\parallel u\parallel . Note that the operator $T:\overline{U}\to {C}_{q}$ is completely continuous (which is known to be compact restricted to bounded sets). From the choice of U, there is no $U\in \partial U$ such that $u=\lambda Tu$ for some $\lambda \in \left(0,1\right)$. Consequently, by Theorem 2.6, we deduce that T has a fixed point $u\in \overline{U}$ which is a solution of problem (1.1). This completes the proof. □

The next existence result is based on the Leray-Schauder continuation theorem.

Theorem 3.6 Let $f:I×{R}^{2}\to R$ be an S-Carathéodory function. Suppose further that there exist functions $p\left(t\right),q\left(t\right),r\left(t\right)\in {L}^{1}\left(I,{R}^{+}\right)$ with $tp\left(t\right)\in {L}^{1}\left(I,{R}^{+}\right)$ such that

$|f\left(t,u,v\right)|\le p\left(t\right)|u|+q\left(t\right)|v|+r\left(t\right),\phantom{\rule{1em}{0ex}}\mathit{\text{for a.e.}}t\in I\mathit{\text{and}}\left(u,v\right)\in {R}^{2}.$

Then BVP (1.1) has at least one solution provided $\left(N+1\right)P+{P}_{1}+Q<1$, where $N=max\left\{\frac{1}{\alpha },1-\frac{1}{\alpha }\right\}$.

Proof We consider the space

$P=\left\{u\in {C}_{q}:{D}_{q}u\left(0\right)=0,{D}_{q}u\left(1\right)=\alpha u\left(1\right)\right\}$

and define the operator ${T}_{1}:P×\left[0,1\right]\to P$ by

${T}_{1}\left(u,\lambda \right)=\lambda Tu=\lambda {\int }_{0}^{1}G\left(t,s;q\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s,\phantom{\rule{1em}{0ex}}t\in I.$
(3.2)

Obviously, we can see that $P\subset {C}_{q}$. In view of Lemma 3.4, it is easy to know that for each $\lambda \in \left[0,1\right]$, ${T}_{1}\left(u,\lambda \right)$ is completely continuous in P. It is clear that $u\in P$ is a solution of BVP (1.1) if and only if u is a fixed point of ${T}_{1}\left(\cdot ,1\right)$. Clearly, ${T}_{1}\left(u,0\right)=0$ for each $u\in P$. If for each $\lambda \in \left[0,1\right]$ the fixed points of ${T}_{1}\left(\cdot ,1\right)$ in P belong to a closed ball of P independent of λ, then the Leray-Schauder continuation theorem completes the proof.

Next we show that the fixed point of ${T}_{1}\left(\cdot ,1\right)$ has a priori bound M, which is independent of λ. Assume that $u={T}_{1}\left(u,\lambda \right)$, and set

$\begin{array}{c}P={\int }_{0}^{1}p\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s,\phantom{\rule{2em}{0ex}}{P}_{1}={\int }_{0}^{1}sp\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s,\hfill \\ Q={\int }_{0}^{1}q\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s,\phantom{\rule{2em}{0ex}}R={\int }_{0}^{1}r\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s.\hfill \end{array}$

By (2.5), it is clear that $|G\left(t,s;q\right)|\le N$ for each $\alpha \ne 0$. For any $u\in P$, we have

$\begin{array}{rl}|u\left(t\right)|& =|u\left(1\right)-{\int }_{t}^{1}{D}_{q}u\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\le |\frac{1}{\alpha }{D}_{q}u\left(1\right)|+|{\int }_{t}^{1}{D}_{q}u\left(s\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le \left(N+1-t\right){\parallel {D}_{q}u\parallel }_{\mathrm{\infty }}\le \left(N+1+t\right){\parallel {D}_{q}u\parallel }_{\mathrm{\infty }},\phantom{\rule{1em}{0ex}}t\in I,\end{array}$

and so it holds that

$\begin{array}{rl}{\parallel {D}_{q}u\parallel }_{\mathrm{\infty }}& \le {\parallel \lambda f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\parallel }_{{L}^{1}}\le {\parallel f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\parallel }_{{L}^{1}}\\ \le {\parallel p\left(t\right)|u\left(s\right)|+q\left(t\right)|{D}_{q}u\left(s\right)|+r\left(s\right)\parallel }_{{L}^{1}}\\ \le \left(\left(N+1\right)P+{P}_{1}+Q\right){\parallel {D}_{q}u\parallel }_{\mathrm{\infty }}+R;\end{array}$

therefore,

${\parallel {D}_{q}u\parallel }_{\mathrm{\infty }}\le \frac{R}{1-\left(\left(N+1\right)P+{P}_{1}+Q\right)}:={M}_{1}.$

At the same time, we have

$\begin{array}{rl}|u\left(t\right)|& \le \lambda |{\int }_{0}^{1}G\left(t,s,q\right)f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s|\\ \le N{\int }_{0}^{1}|f\left(s,u\left(s\right),{D}_{q}u\left(s\right)\right)|\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ \le N{\int }_{0}^{1}\left(p\left(t\right)|u\left(s\right)|+q\left(t\right)|{D}_{q}u\left(s\right)|+r\left(s\right)\right)\phantom{\rule{0.2em}{0ex}}{d}_{q}s\\ \le NP{\parallel u\parallel }_{\mathrm{\infty }}+NQ{\parallel {D}_{q}u\parallel }_{\mathrm{\infty }}+NR,\phantom{\rule{1em}{0ex}}t\in I,\end{array}$

and so

${\parallel u\parallel }_{\mathrm{\infty }}\le \frac{N\left(Q{M}_{1}+R\right)}{1-NP}:={M}_{2}.$

Set $M=max\left\{{M}_{1},{M}_{2}\right\}$, which is independent of λ. So, BVP (1.1) has at least one solution. This completes the proof. □

## 4 Example

Example 4.1 Consider the following BVP:

$\left\{\begin{array}{l}{D}_{q}^{2}u\left(t\right)={e}^{t}+\frac{1}{4}sin\left(u\left(t\right)\right)+\frac{1}{8}{tan}^{-1}\left({D}_{q}u\left(t\right)\right),\phantom{\rule{1em}{0ex}}t\in I,\\ {D}_{q}u\left(0\right)=0,\phantom{\rule{2em}{0ex}}{D}_{q}u\left(1\right)=\frac{1}{2}u\left(1\right).\end{array}$
(4.1)

Here, $f\left(t,u\left(t\right),{D}_{q}u\left(t\right)\right)={e}^{t}+\frac{1}{4}sin\left(u\left(t\right)\right)+\frac{1}{8}{tan}^{-1}\left({D}_{q}u\left(t\right)\right)$, $q=\frac{1}{2}$, $\alpha =\frac{1}{2}$. Clearly, $|f\left(t,{u}_{1},{v}_{1}\right)-f\left(t,{u}_{2},{v}_{2}\right)|\le \frac{1}{4}|{u}_{1}-{u}_{2}|+\frac{1}{8}|{v}_{1}-{v}_{2}|$. Then ${L}_{1}=\frac{1}{4}$, ${L}_{2}=\frac{1}{8}$ and ${L}_{1}+{L}_{2}=\frac{3}{8}<\alpha$. By Corollary 3.2, we obtain that BVP (4.1) has a unique solution.

Example 4.2 Consider the following BVP:

$\left\{\begin{array}{l}{D}_{q}^{2}u\left(t\right)=\frac{t}{20}sin\left(u\left(t\right)\right)+\frac{1}{\left(2+\sqrt{2}\right)\sqrt{t}}{\left({D}_{q}u\left(t\right)\right)}^{\beta }+\frac{3}{2}t,\phantom{\rule{1em}{0ex}}t\in I,\\ {D}_{q}u\left(0\right)=0,\phantom{\rule{2em}{0ex}}{D}_{q}u\left(1\right)=\frac{1}{2}u\left(1\right).\end{array}$
(4.2)

Here, $q=\frac{1}{2}$, $0<\beta <1$. It is obvious that $|f\left(t,u,v\right)|\le \frac{t}{20}|u|+\frac{1}{\left(2+\sqrt{2}\right)\sqrt{t}}|v|+\frac{3}{2}t$, where $p\left(t\right)=\frac{t}{20}$, $q\left(t\right)=\frac{1}{\left(2+\sqrt{2}\right)\sqrt{t}}$, $r\left(t\right)=\frac{3}{2}t$. Then $P=\frac{1}{30}$, ${P}_{1}=\frac{1}{35}$, $Q=\frac{1}{2}$, $R=1$, so $\left(N+1\right)P+{P}_{1}+Q=\frac{22}{35}<1$. By Theorem 3.6, we obtain that BVP (4.2) has at least one solution.

## References

1. Jackson FH: On q -difference equations. Am. J. Math. 1910, 32: 305–314. 10.2307/2370183

2. Carmichael RD: The general theory of linear q -difference equations. Am. J. Math. 1912, 34: 147–168. 10.2307/2369887

3. Mason TE: On properties of the solutions of linear q -difference equations with entire function coefficients. Am. J. Math. 1915, 37: 439–444. 10.2307/2370216

4. Adams CR: On the linear ordinary q -difference equation. Ann. Math. 1928, 30: 195–205. 10.2307/1968274

5. Strominger A: Information in black hole radiation. Phys. Rev. Lett. 1993, 71: 3743–3746. 10.1103/PhysRevLett.71.3743

6. Youm D: q -deformed conformal quantum mechanics. Phys. Rev. D 2000., 62: Article ID 095009

7. Lavagno A, Swamy PN: q -Deformed structures and nonextensive statistics: a comparative study. Physica A 2002, 305(1–2):310–315. Non extensive thermodynamics and physical applications (Villasimius, 2001) 10.1016/S0378-4371(01)00680-X

8. Ahmad B: Boundary value problems for nonlinear third-order q -difference equations. Electron. J. Differ. Equ. 2011, 94: 1–7.

9. Ahmad B, Nieto JJ: On nonlocal boundary value problem of nonlinear q -difference equations. Adv. Differ. Equ. 2012., 2012: Article ID 81. doi:10.1186/1687–1847–2012–81

10. Ahmad B, Alsaedi A, Ntouyas SK: A study of second-order q -difference equations with boundary conditions. Adv. Differ. Equ. 2012., 2012: Article ID 35. doi:10.1186/1687–1847–2012–35

11. El-Shahed M, Hassan HA: Positive solutions of q -difference equation. Proc. Am. Math. Soc. 2010, 138: 1733–1738.

12. Ahmad B, Ntouyas SK: Boundary value problems for q -difference inclusions. Abstr. Appl. Anal. 2011., 2011: Article ID 292860

13. Ahmad B, Nieto JJ: Basic theory of nonlinear third-order q -difference equations and inclusions. Math. Model. Anal. 2013, 18(1):122–135. 10.3846/13926292.2013.760012

14. Gasper G, Rahman M: Basic Hypergeometric Series. Cambridge University Press, Cambridge; 1990.

15. Kac V, Cheung P: Quantum Calculus. Springer, New York; 2002.

16. Granas A, Dugundji J: Fixed Point Theory. Springer, New York; 2005.

## Acknowledgements

Research supported by the Natural Science Foundation of China (10901045), (11201112), the Natural Science Foundation of Hebei Province (A2009000664), (A2011208012) and the Foundation of Hebei University of Science and Technology (XL201047), (XL200757).

## Author information

Authors

### Corresponding author

Correspondence to Changlong Yu.

### Competing interests

The authors declare that they have no competing interests.

### Authors’ contributions

Each of the authors, CLY and JFW contributed to each part of this work equally and read and approved the final version of the manuscript.

## Rights and permissions

Reprints and Permissions

Yu, C., Wang, J. Existence of solutions for nonlinear second-order q-difference equations with first-order q-derivatives. Adv Differ Equ 2013, 124 (2013). https://doi.org/10.1186/1687-1847-2013-124

• Accepted:

• Published:

• DOI: https://doi.org/10.1186/1687-1847-2013-124

### Keywords

• q-difference equations
• Leray-Schauder nonlinear alternative
• boundary value problem
• fixed point theorem 