On nonlocal fractional q-integral boundary value problems of fractional q-difference and fractional q-integrodifference equations involving different numbers of order and q

In this paper, we study some new class of nonlocal three-point fractional q-integral boundary value problems of a nonlinear fractional q-difference equation and a nonlinear fractional q-integrodifference equation. Our problems contain different numbers of order and q in derivatives and integrals. The existence and uniqueness results are based on Banach’s contraction mapping principle and Krasnoselskii’s fixed point theorem. In addition, some examples are presented to illustrate the importance of these results.

Jackson [1] initiated quantum calculus or q-difference calculus that can describe many phenomena in various fields of science and engineering. Basic definitions and properties of q-difference calculus can be found in the book [2]. For the fractional q-difference calculus originating with Al-Salam [3] and Agarwal [4], we refer to the book of Annaby and Mansour [5].

A class of integral boundary value problems appeared in different areas of applied mathematics and physics. For instance, blood flow problems, population dynamics, heat conduction, underground water flow, thermo-elasticity, plasma physics, chemical engineering and so on can be reduced to nonlocal integral boundary problems. For comments on the importance of integral boundary problems, we refer the reader to the papers by Webb and Infante [6, 7], Gallardo [8], Karakostas and Tsamatos [9], Lomtatidze and Malaguti [10], and the references therein. For information as regards the theory of integral equations and their applications to integral boundary value problems, we refer to the books of Agarwal and O’Regan [11] and Corduneanu [12].

For some recent works on q-integral boundary value problems, we refer to [13–29] and the references cited therein. For example, Ahmad et al. [21] considered the following boundary value problem of nonlinear fractional q-difference equation with nonlocal and sub-strip type boundary conditions:

where \(1<\nu\leq2\), \(0< q<1\), \(0<\xi<\eta<1\), \(^{C}D_{q}^{\nu}\) is the Caputo fractional q-derivative of order ν, \(f\in C([0,1]\times\mathbb {R},\mathbb{R})\), and \(g\in C([0,1],\mathbb{R})\) are given functions. The existence results for the problem (1.1) are shown by applying Banach’s contraction mapping principle and a fixed point theorem due to O’Regan.

Almeida and Martins [26] proposed the following fractional q-difference equation with three-point integral boundary conditions:

where \(2<\alpha<3\), \(0<\eta\leq1\), \(^{C}D_{q}^{\alpha}\) is the Caputo fractional q-derivative of order α, \(g\in C([0, 1]\times \mathbb{R}, \mathbb{R})\). They presented the existence and uniqueness results for the problem (1.2) by employing Banach’s contraction mapping principle, Krasnoselskii fixed point theorem and Leray-Schauder alternative.

Presently, there is a development of boundary value problems for fractional q-difference equations showing an operation of the investigative function. The study may also have another function, related to our interest. These creations are incorporating nonlocal conditions that are both extensive and more complex.

The results mentioned above are the motivation for this research. In this article, we discuss the existence and uniqueness results of solutions to a nonlinear fractional q-difference equation with nonlocal three-point fractional q-integral boundary conditions of the form

and a nonlinear fractional q-integrodifference equation with nonlocal three-point fractional q-integral boundary conditions of the form

where \(p,q,w \in(0,1)\), \(\alpha\in(1,2]\), \(\nu\in(0,1]\), \(\beta,\gamma >0\), and \(\eta\in(0,T)\) are given constants, \(D_{q}^{\alpha}\) and \(D_{w}^{\nu}\) are the Riemann-Liouville fractional q-derivative of order α and w-derivative of order ν, respectively, \(f\in C([0,T]\times \mathbb{R}\times\mathbb{R},\mathbb{R})\) and \(g \in C([0, T],\mathbb {R^{+}})\) are given functions, \(\rho\in C([0,T],{\mathbb{R}})\rightarrow {\mathbb{R}}\) is a given functional, and for \(\varphi\in C([0, T]\times [0, T],[0,\infty))\),

The plan of this paper is as follows. In Section 2, we recall some definitions and basic lemmas. In Sections 3 and 4, we prove the existence and uniqueness results for the boundary value problems (1.3) and (1.4) by employing Banach’s contraction mapping principle and Krasnoselskii’s fixed point theorem. Some illustrative examples are presented in the last section.

In the following, there are notations, definitions, and lemmas which are used in the main results. Let \(q\in(0,1)\) and define

The q-analog of the power function \((a-b)^{(n)}\) with \(n\in\mathbb {N}_{0}:=[0,1,2,\ldots]\) is

More generally, if \(\alpha\in\mathbb{R}\), then

Note that, if \(b = 0\) then \(a^{(\alpha)} = a^{\alpha}\). We also use the notation \(0^{(\alpha)}=0\) for \(\alpha>0\). The q-gamma function is defined by

and satisfies \(\Gamma_{q}(x+1)=[x]_{q}\Gamma_{q}(x)\).

Remark

[13]

We note that if \(\alpha>0\) and \(a\leq b\leq t\), then \((t-a)^{(\alpha)}\geq(t-b)^{(\alpha)}\).

Definition 2.1

[4]

For \(\alpha\geq0\) and f defined on \([0, T]\), the fractional q-integral of the Riemann-Liouville type is defined by

and \((I^{0}_{q} f)(x) = f(x)\).

Definition 2.2

[15]

For \(\alpha\geq0\) and f defined on \([0, T]\), the fractional q-derivative of the Riemann-Liouville type of order α is defined by

and \((D^{0} _{q}f)(x) = f(x)\), where m is the smallest integer that is greater than or equal to α.

Definition 2.3

[2]

For any \(x,s>0\),

is called the q-beta function.

Lemma 2.1

[4]

Let \(\alpha,\beta\geq0\) and f be a function defined on \([0, T]\). Then the next formulas hold:

Lemma 2.2

[15]

Let \(\alpha> 0\) and N be a positive integer. Then the following equality holds:

Lemma 2.3

[16]

Let \(\alpha,\beta\geq0\) and \(0< p,q<1\). Then the following formulas hold:

To define the solution of the boundary value problems (1.3) and (1.4), we need the following lemma, which deals with a linear variant of the boundary value problems (1.3) and (1.4) and gives a representation of the solution.

Lemma 2.4

Let \(p,q\in(0,1)\), \(\alpha\in(1,2 ]\), \(\beta>0\), \(\eta\in(0,T)\), functions \(y \in C([0,T],\mathbb{R})\) and \(g \in C([0,T],\mathbb {R}^{+})\), and a functional \(\rho: C([0,T],\mathbb{R})\rightarrow\mathbb {R}\). Then the boundary value problem

is equivalent to the integral equation

Proof

Consider \(m=2\). By Definition 2.2 and Lemma 2.1, we obtain

and

The first condition of (2.2) implies

Taking the fractional p-integral of order \(\beta>0\) for (2.6) and the second condition of (2.2), we get

Solving the system of linear equations (2.6) and (2.7), for the unknown constants \(C_{1}\) and \(C_{2}\), we have

Substituting the constants \(C_{1}\) and \(C_{2}\) into (2.5), we obtain (2.3).

This completes the proof. □

Let \(\mathcal{C}_{\mathcal{A}} = C([0,T],\mathbb{R})\) be a Banach space of all continuous functions from \([0,T]\) to \(\mathbb{R}\), endowed with the norm defined by

where \(\|x\|= \sup_{t\in[0,T]} |x(t)|\) and \(\|D_{\omega}^{\nu}x\|= \sup_{t\in[0,T]} |D_{\omega}^{\nu}x(t)|\). Define the operator \(\mathcal{A}:\mathcal{C}_{\mathcal{A}}\rightarrow\mathcal {C}_{\mathcal{A}}\) by

Observe that the problem (1.3) has solutions if and only if the operator \(\mathcal{A}\) has fixed points.

Now, we are in the position to establish the main results. Our first result is based on Banach’s fixed point theorem.

Theorem 3.1

Assume that a functional \(\rho\in C([0,T],\mathbb{R})\rightarrow \mathbb{R}\), \(f:[0,T]\times\mathbb{R}\times\mathbb{R}\rightarrow\mathbb {R}\) and \(g:[0,T]\rightarrow\mathbb{R}\) are continuous functions satisfying the following conditions:

(H₁):

There exist positive numbers \(L_{1}\), \(L_{2}\) such that, for each \(t\in[0,T]\) and \(x,y\in{\mathcal{C}}_{\mathcal{A}}\),

$$\bigl|f\bigl(t,x,D_{w}^{\nu}x\bigr)-f\bigl(t,y,D_{w}^{\nu}y\bigr)\bigr|\leq L_{1}|x-y|+L_{2}\bigl|D_{w}^{\nu}x-D_{w}^{\nu}y\bigr|. $$

(H₂):

There exists a positive number τ such that, for each \(x,y\in{\mathcal{C}}_{\mathcal{A}}\),

$$\bigl|\rho(x)-\rho(y)\bigr|\leq\tau\|x-y\|_{\mathcal{C_{A}}}. $$

(H₃):

For each \(t\in[0,T]\), \(0< n< g(t)< N\).

(H₄):

\(\Theta:=\lambda (\Omega+\Lambda )+\frac{\tau\Gamma _{q}(\alpha+1)}{\eta^{2}}\Lambda<1\),

Then the boundary value problem (1.3) has a unique solution.

Proof

We transform the boundary value problem (1.3) into a fixed point problem \(x=\mathcal{A}x\), where \(\mathcal{A}:\mathcal{C}_{\mathcal{A}}\rightarrow\mathcal{C}_{\mathcal {A}}\) is defined by (3.1). Assuming that \(\sup_{t\in[0,T]}{|f(t,0,0)|} = M\) and \(\sup_{x\in{\mathcal{C}}_{\mathcal{A}}}{|\rho(x)|} = K\), we choose a constant R satisfied with

Now, we will show that \(\mathcal{A}B_{R}\subset B_{R}\), where \(B_{R} = \{x \in \mathcal{C}_{\mathcal{A}}: \|x\|_{\mathcal{C_{A}}} \leq R\}\). For all \(x, y\in\mathcal{C}_{\mathcal{A}}\) and for each \(t\in[0,T]\), we have

and

Therefore, we obtain \(\|{\mathcal{A}}x\|_{\mathcal{C_{A}}}\leq R\) and hence \(\mathcal{A}B_{R}\subset B_{R}\).

Next, we will show that \(\mathcal{A}\) is a contraction. Denote

For all \(x, y\in\mathcal{C}_{\mathcal{A}}\) and for each \(t\in[0,T]\), we have

and

Thus, \(\|{\mathcal{A}}x-{\mathcal{A}}y\|_{\mathcal{C_{A}}}\leq \Theta\|x-y\|_{\mathcal{C_{A}}}\). From (H₄), we see that \(\mathcal{A}\) is a contraction.

Hence, the conclusion of the theorem follows by Banach’s contraction mapping principle. This completes the proof. □

Our second result is based on the following Krasnoselskii fixed point theorem.

Theorem 3.2

(Krasnoselskii fixed point theorem) [30]

Let K be a bounded closed convex and nonempty subset of a Banach space X. Let A, B be operators such that

1. (i)

   \(Ax+By \in K\) whenever \(x,y \in K\),

2. (ii)

   A is compact and continuous,

3. (iii)

   B is a contraction mapping.

Then there exists \(z \in K\) such that \(z = Az+Bz\).

Theorem 3.3

(Arzela-Ascoli theorem) [30]

Let \(D\subseteq\mathbb{R}^{n}\) be a bounded domain, \(K\subseteq C(\overline{D},\mathbb{R})\) be bounded and the following property of equicontinuity holds. For every \(\epsilon>0\), there exists \(\delta>0\), so that

Then K̅ is compact.

Theorem 3.4

Assume that (H₂)-(H₃) hold. In addition, \(f:[0,T]\times\mathbb {R}\times\mathbb{R}\rightarrow\mathbb{R}\) is a continuous function satisfying the following condition:

(H₅):

For all \((t,x,D_{w}^{\nu}x)\in[0,T]\times \mathbb{R}\times \mathbb{R}\), with \(\mu\in C([0,T], \mathbb{R}^{+})\),

$$\bigl|f\bigl(t,x,D_{w}^{\nu}x\bigr)\bigr|\leq\mu(t). $$

then the boundary value problem (1.3) has at least one solution on \([0,T]\).

Proof

Set \(\sup_{t\in[0,T]}|\mu(t)|=\| \mu\|\), and choose a constant

In view of Lemma 2.4, we define the operators \(\mathcal{A}_{1}\) and \(\mathcal{A}_{2}\) on the ball \(B_{R}=\{x\in\mathcal {C}_{\mathcal{A}}:\|x\|_{\mathcal{C_{A}}}\leq R\}\) by

For all \(x, y \in B_{R}\), by computing directly, we have

Similarly to the proof above and Theorem 3.1, we obtain \(\|D_{w}^{\nu}\mathcal{A}_{1}x+D_{w}^{\nu}\mathcal{A}_{2}y \|< R\), and hence \(\|\mathcal{A}_{1}x+\mathcal{A}_{2}y\|_{\mathcal{C_{A}}}< R\). Therefore,

The condition (3.3) implies that \(\mathcal{A}_{2}\) is a contraction mapping.

Next, we will show that \(\mathcal{A}_{1}\) is compact and continuous. Continuity of f coupled with the assumption (H₄) implies that the operator \(\mathcal{A}_{1}\) is continuous and uniformly bounded on \(B_{R}\). For \(t_{1}, t_{2} \in[0,T]\) with \(t_{1}< t_{2}\), we have

Similarly to the proof above and Theorem 3.1, we obtain

Actually, as \(|t_{2}-t_{1}|\rightarrow0\), the right-hand side of the above inequality tends to be zero. So \(\mathcal{A}_{1}\) is relatively compact on \(B_{R}\). Hence, by the Arzela-Ascoli theorem, \(\mathcal{A}_{1}\) is compact on \(B_{R}\).

Therefore, all the assumptions of Theorem (3.2) are satisfied and the conclusion of Theorem 3.2 implies that the boundary value problem (1.3) has at least one solution on \([0,T]\). This completes the proof. □

Let \(\mathcal{C} = C([0,T],\mathbb{R})\) be a Banach space of all continuous functions from \([0,T]\) to \(\mathbb{R}\) endowed with the norm \(\|x\|= \sup_{t\in[0,T]} |x(t)|\). Define the operator \(\mathcal{B}:\mathcal{C}\rightarrow\mathcal{C}\) as follows:

Observe that the problem (1.4) has solutions if and only if the operator \(\mathcal{B}\) has fixed points.

Our first result is based on Banach’s fixed point theorem.

Theorem 4.1

Assume that a functional \(\rho: C([0,T],\mathbb{R})\rightarrow\mathbb {R}\) is continuous, \(f:[0,T]\times\mathbb{R}\times\mathbb{R}\rightarrow \mathbb{R}\) is continuous and maps bounded subsets of \([0,T]\times\mathbb{R}\times\mathbb{R}\) into relatively compact subsets of \(\mathbb{R}\), \(g:[0,T]\rightarrow\mathbb{R^{+}}\), and \(\varphi: [0, T]\times[0, T]\rightarrow[0,\infty)\) are continuous functions. Let \(\varphi_{0}=\sup_{(t,s)\in[0, T]\times[0, T]}\{\varphi(t,s)\}\) and (H₃) hold. In addition, ρ and f satisfy the following conditions:

(H₆):

There exists a positive number τ such that, for all \(x,y\in{\mathcal{C}}\),

$$\bigl|\rho(x)-\rho(y)\bigr|\leq\tau\|x-y\|. $$

(H₇):

There exist positive numbers \(l_{1}\), \(l_{2}\) such that, for each \(t\in[0,T]\) and \(x,y\in{\mathcal{C}}\),

$$\bigl|f\bigl(t,x,\Psi_{w}^{\gamma}x\bigr)-f\bigl(t,y, \Psi_{w}^{\gamma}y\bigr)\bigr|\leq l_{1}|x-y|+l_{2}\bigl| \Psi _{w}^{\gamma}x-\Psi_{w}^{\gamma}y\bigr|. $$

(H₈):

\(\Upsilon:= (l_{1}+l_{2} \frac{\varphi_{0} T^{\gamma}}{\Gamma_{w}(\gamma+1)} ) \Omega+\frac{\tau\Gamma_{q}(\alpha+1)}{\eta ^{2}}\Lambda<1\), where Ω, Λ are defined as (3.2).

Proof

We transform the boundary value problem (1.4) into a fixed point problem \(x=\mathcal{B}x\), where \(\mathcal{B}:\mathcal{C}\rightarrow\mathcal{C}\) is defined by (4.1). Assuming that \(\sup_{t\in[0,T]}{|f(t,0,0)|} = M\) and \(\sup_{x\in{\mathcal{C}}}{\rho(x)} = K\), we choose a constant ρ satisfied with

Now, we will show that \(\mathcal{B}B_{\rho}\subset B_{\rho}\), where \(B_{\rho}= \{x \in \mathcal{C}: \|x\| \leq\rho\}\). For all \(x\in B_{\rho}\), we have

Therefore, \({\mathcal{B}}B_{\rho}\subset B_{\rho}\).

Next, we will show that \(\mathcal{B}\) is a contraction. For all \(x, y\in \mathcal{C}\) and for each \(t\in[0,T]\), we have

By (H₈), we have \(\mathcal{B}\) is a contraction. Thus, the conclusion of the theorem follows by Banach’s contraction mapping principle. This completes the proof. □

Our second result is based on the following Krasnoselskii’s fixed point theorem.

Theorem 4.2

Assume that (H₃) and (H₆) hold. In addition, \(f:[0,T]\times\mathbb {R}\times\mathbb{R}\rightarrow\mathbb{R}\) is a continuous function satisfying the following condition:

(H₉):

For all \((t,x,\Psi_{w}^{\gamma}x)\in[0,T]\times \mathbb{R}\times\mathbb{R}\), with \(\sigma\in C([0,T], \mathbb{R}^{+})\),

$$\bigl|f\bigl(t,x,\Psi_{w}^{\gamma}x\bigr)\bigr|\leq\sigma(t). $$

then the boundary value problem (1.4) has at least one solution on \([0,T]\).

Proof

Set \(\sup_{t\in[0,T]}|\sigma (t)|=\|\sigma\|\) and choose a constant

In view of Lemma 2.4, we define the operators \({\mathcal{B}}_{1}\) and \({\mathcal{B}}_{2}\) on the ball \(B_{\rho}=\{x\in \mathcal{C}:\|x\|\leq\rho\}\) by

For \(x, y \in B_{\rho}\), by computing directly, we have

Therefore \(\mathcal{B}_{1}x+\mathcal{B}_{2}y\in B_{\rho}\). The condition (4.3) implies that \(\mathcal{B}_{2}\) is a contraction mapping.

Next, we will show that \(\mathcal{B}_{1}\) is compact and continuous. Continuity of f coupled with the assumption (H₇) implies that the operator \(\mathcal{B}_{1}\) is continuous and uniformly bounded on \(B_{\rho}\). For \(t_{1}, t_{2} \in[0,T]\) with \(t_{1}< t_{2}\), we have

Actually, as \(|t_{2}-t_{1}|\rightarrow0\), the right-hand side of the above inequality tends to be zero. So \(\mathcal{B}_{1}\) is relatively compact on \(B_{\rho}\). Hence, by the Arzela-Ascoli theorem, \(\mathcal{B}_{1}\) is compact on \(B_{\rho}\).

Therefore, all the assumptions of Theorem 4.2 are satisfied and the conclusion of Theorem 4.2 implies that the boundary value problem (1.4) has at least one solution on \([0,T]\). This completes the proof. □

In this section, we give some examples to illustrate our results.

Example 5.1

Consider the following fractional q-integral boundary value problem:

where \(0< t_{1},t_{2},\ldots,t_{n}<1\) and \(C_{i}\) are given positive constants with \(\sum_{i=1}^{n}C_{i}<\frac{1}{400}\).

Here \(\alpha=\frac{4}{3}\), \(\beta=\frac{5}{2}\), \(\eta=\frac{1}{4}\), \(\nu =\frac{2}{3}\), \(q=\frac{1}{2}\), \(p=\frac{1}{3}\), \(w=\frac{1}{4}\), \(T=1\), \(g(t)=e^{\sin(2\pi t)}\), and \(f(t,x)=\frac{e^{-\cos^{2}(2\pi t)}}{20+e^{\sin^{2}(2\pi t)}}\cdot\frac{|x(t)|+|D_{\frac{1}{4}}^{\frac {2}{3}} x(t)|}{1+|x(t)|}\).

Since \(|f(t,x,D_{\frac{1}{2}}^{\frac{4}{3}} x)-f(t,y,D_{\frac {1}{2}}^{\frac{4}{3}} y)|\leq\frac{1}{101}|x-y|+\frac{1}{101}|D_{\frac {1}{2}}^{\frac{4}{3}} x-D_{\frac{1}{2}}^{\frac{4}{3}} y|\), (H₁) is satisfied with \(L_{1}=L_{2}=\frac{1}{101}\), so \(\lambda=\frac{2}{101}\).

Also, we get \(|\rho(x)-\rho(y)|=\vert \sum_{i=1}^{n}C_{i}x(t_{i})-\sum_{i=1}^{n}C_{i}y(t_{i})\vert \leq\sum_{i=1}^{n}C_{i}|x-y|\). So, (H₂) holds with \(\tau=\sum_{i=1}^{n}C_{i}<\frac{1}{400}\).

Since \(\frac{1}{e}< g(t)< e\), (H₃) is satisfied with \(N=e\), \(n=\frac {1}{e}\).

We can show that

Therefore, we get

Hence, by Theorem 3.1, problem (5.1) has a unique solution on \([0,1]\).

Example 5.2

Consider the following fractional q-integral boundary value problem:

where \(0< t_{1},t_{2},\ldots,t_{n}<1\) and \(C_{i}\) are given positive constants with \(\sum_{i=1}^{n}C_{i}<\frac{1}{500}\).

Here \(\alpha=\frac{4}{3}\), \(\beta=\frac{5}{2}\), \(\eta=\frac{1}{4}\), \(\gamma=\frac{8}{3}\), \(q=\frac{1}{2}\), \(p=\frac{1}{3}\), \(w=\frac {1}{4}\), \(T=1\), \(g(t)=e^{\sin(2\pi t)}\), \(f(t,x)=\frac{e^{-\cos^{2}(2\pi t)}}{20+e^{\sin^{2}(2\pi t)}}\cdot\frac {|x(t)|+|\Psi_{\frac{1}{4}}^{\frac{8}{3}} x(t)|}{1+|x(t)|}\), \(\Psi _{\frac{1}{4}}^{\frac{8}{3}} x(t)=\frac{1}{\Gamma_{\frac{1}{4}}(\frac {8}{3})} \int_{0}^{t} (t-\frac{s}{4} )^{(\frac {5}{3})}\frac{e^{-(s-t)}}{10} x(s) \,d_{w}s \) and \(\varphi(t,s)=\frac{e^{-(s-t)}}{10}\).

Also, we get \(|\rho(x)-\rho(y)|=\vert \sum_{i=1}^{n}C_{i}x(t_{i})-\sum_{i=1}^{n}C_{i}y(t_{i})\vert \leq\sum_{i=1}^{n}C_{i}|x-y|\). So, (H₂) holds with \(\tau=\sum_{i=1}^{n}C_{i}<\frac{1}{500}\).

Since \(|f(t,x,\Psi_{\frac{1}{4}}^{\frac{8}{3}} x)-f(t,y,\Psi_{\frac {1}{4}}^{\frac{8}{3}} y)|\leq\frac{1}{21}|x-y|+\frac{1}{21}|\Psi _{\frac{1}{4}}^{\frac{8}{3}} x-\Psi_{\frac{1}{4}}^{\frac{8}{3}} y| \) and \(\varphi_{0}=\frac{e}{10}\), (H₅) is satisfied with \(l_{1}=l_{2}=\frac {1}{21}\). So \(l_{1}+\frac{e l_{2}}{10 (\frac{8}{3}+1)}=0.0977\).

By Example 5.1, we get \(\Omega\approx 2.2358 \) and \(\Lambda \approx5.0088\). Therefore, we have

Hence, by Theorem 4.1, problem (5.2) has a unique solution on \([0,1]\).

The author is highly grateful to the editors and the referees for their valuable suggestions, which improved the article. This research was funded by King Mongkut’s University of Technology North Bangkok. Contract no. KMUTNB-GEN-58-13.

Affiliations

1. Nonlinear Dynamic Analysis Research Center, Department of Mathematics, Faculty of Applied Science, King Mongkut’s University of Technology North Bangkok, Bangkok, Thailand

   • Thanin Sitthiwirattham

Corresponding author

Correspondence to Thanin Sitthiwirattham.

Competing interests

The author declares to have no competing interests.

Reprints and Permissions