Applied Mathematics Letters 98 (2019) 292–299

Contents lists available at ScienceDirect

Applied Mathematics Letters

www.elsevier.com/locate/aml

A difference method for solving the q-fractional differential

equations
Tie Zhang, Yongchao Tang ∗
Department of Mathematics and the State Key Laboratory of Synthetical Automation for Process
Industries, Northeastern University, Shenyang 110819, PR China

article info abstract

Article history: In this article, we study the Caputo type q-fractional differential equation:
Received 13 May 2019 c D α x(t) = f (t, x(t)), 0 < α, q < 1. By using the piecewise linear interpolation
q
Received in revised form 20 June 2019 approximation, we present a difference formula for discretizing the Caputo type
Accepted 20 June 2019 q-fractional derivative c Dqα x(t) and derive the truncation error boundness. This
Available online 27 June 2019 difference formula is used to solve the initial value problem of q-fractional
Keywords: differential equation. Numerical examples are provided to verify the effectiveness
q-fractional derivative of our difference method.
q-fractional integral © 2019 Elsevier Ltd. All rights reserved.
q-fractional differential equation
Difference formula
Truncation error boundness

1. Introduction

In recent years, q-calculus [1–4] and q-fractional calculus [5–9] has attracted much attention in quantum
mechanics and stochastic analysis fields. The Caputo type q-fractional calculus is one of the classical
q-fractional calculuses. The existence of solutions for the Caputo type q-fractional boundary value problems
has been studied by many researches, see [10–13]. By using Lyapunov’s direct method, Jarad [14] studied
the stability of Caputo type q-fractional non-autonomous systems. Wang [15] investigated the existence of
extremal solutions of the Caputo type q-fractional differential equation, while Abdeljawad [16] also gave a
class of solutions by means of a generalized type of q-Mittag-Leffler function. On the numerical methods,
Abdeljawad [17] first presented the successive approximation method to obtain an approximate solution of
the Caputo type q-fractional differential equation. Then, Salahshour [18] investigated the convergence of the
successive approximation method presented by Abdeljawad. Wu [19] used the variational iteration method
and the Lagrange multipliers method to study the Caputo type q-fractional initial value problem. Recently,
Tang [20] gave the existence theorem of solution for the Caputo type q-fractional differential equation by
using the q-beta function and established an iteration method for solving this equation.

∗ Corresponding author.
E-mail addresses: [email protected] (T. Zhang), [email protected] (Y. Tang).

https://doi.org/10.1016/j.aml.2019.06.020
0893-9659/© 2019 Elsevier Ltd. All rights reserved.
 T. Zhang and Y. Tang / Applied Mathematics Letters 98 (2019) 292–299 293

In this article, we consider the Caputo type q-fractional differential equation: c Dqα x(t) = f (t, x(t)),
0 < α, q < 1 with given initial value. We first present a difference formula to discretize the derivative
c α
Dq x(t). This difference formula is constructed by using the piecewise linear interpolation to approximate
the integrand function. We derive the truncation error boundness for this difference formula. Then, using
this difference formula, we establish a difference method for solving the q-fractional differential equation. The
effectiveness of this difference method is verified by numerical experiment. To the authors’ best knowledge,
it is the first time that a difference method is presented for solving the q-fractional differential equations.

2. The q-calculus and q-fractional calculus

The q-analogy of (t − s)α is defined by

∞
∏ t − qi s
(t − s)(α) = tα . (2.1)
i=0
t − q α+i s

If α > 0, t ≥ 0 and a ≤ b ≤ t, it holds that (t − a)(α) ≥ (t − b)(α) [21].

Let 0 < |q| < 1 and α ∈ C\ {−n : n ∈ N ∪ {0}}, where C is the complex set and N = {1, 2, 3, . . .}. Denote
by Γq (α) the q-gamma function [22]

Γq (α) = (1 − q)1−α (1 − q)(α−1) . (2.2)

By a straightforward computation, we know that Γq (α) satisfies the following equation:

Γq (α + 1) = [α]Γq (α), Γq (1) = 1, (2.3)

where [α] = (1 − q α )/(1 − q).

Let q ∈ R be a fixed point, a subset A of C is called q-geometric if qt ∈ A whenever t ∈ A. A typical subset
of A is called the time scale Tq defined by Tq = {q n : n ∈ Z} ∪ {0}, where 0 < q < 1, Z = {0, ±1, ±2, . . .}.

Definition 2.1 ([1]). Let f be a real or complex valued function, t ∈ A, |q| = ̸ 1. The q-derivative is defined
by
dq f (qt) − f (t)
Dq f (t) = f (t) = , t ∈ A\ {0} , (2.4)
dq t (q − 1)t
and
dq f (tq n ) − f (0)
Dq f (0) = f (t)|t=0 = lim , |q| < 1.
dq t n→∞ tq n
By a straightforward computation, the q-derivative of the q-shifted factorial with respect to t is as
follows [23]
Dq (t − s)(α) = [α](t − s)(α−1) . (2.5)

Definition 2.2 ([24]). The q-integral on interval (a, b) is defined by

∫ b ∫ b ∫ a
f (t)dq t = f (t)dq t − f (t)dq t, a, b ∈ A, (2.6)
a 0 0

where ∫ x ∞
∑
f (t)dq t = (1 − q) xq n f (xq n ), x ∈ A. (2.7)
0 n=0
294 T. Zhang and Y. Tang / Applied Mathematics Letters 98 (2019) 292–299

Lemma 2.1 ([17]). Let 0 < q < 1 and f be a function defined on A containing zero point. For two points
a, t ∈ A, t > a, we have for the q-derivative operator Dq with respect to t,
∫ t ∫ t
Dq f (t, s)dq s = Dq f (t, s)dq s + f (qt, t). (2.8)
a a

Definition 2.3 ([23]). Let t ∈ A, α ̸= −1, −2, . . . and a ≥ 0. The α-order q-fractional integral of the
0
Riemann–Liouville type with the lower limit point a is defined by Iq,a f (t) = f (t) and
∫ t
α 1
Iq,a f (t) = (t − qs)(α−1) f (s)dq s. (2.9)
Γq (α) a

Definition 2.4 ([25]). Let a ∈ Tq . The α-order Riemann–Liouville type q-fractional derivative of a function
f (t) : (a, ∞) → R is defined by
−α
{
(Iq,a f )(t), α ≤ 0,
α
(Dq,a f )(t) = ⌈α⌉ ⌈α⌉−α
(2.10)
(Dq,a Iq,a f )(t), α > 0,
where ⌈α⌉ denotes the smallest integer that is greater or equal to α.

Definition 2.5 ([25],[17]). Let a ∈ Tq . The α-order Caputo type q-fractional derivative of a function
f (t) : (a, ∞) → R is defined by
−α
{
(Iq,a f )(t), α ≤ 0,
c α
( Dq,a f )(t) = ⌈α⌉−α ⌈α⌉
(2.11)
(Iq,a Dq,a f (t)), α > 0.
α
For simplicity, we often use the notation Iqα f (t) instead of Iq,0 f (t), the notation Dqα f (t) instead of
α c α c α
Dq,0 the notation Dq f (t) instead of Dq,0 f (t) respectively.
f (t),

Definition 2.6 ([5]). A function f defined on [0, b] is called q-absolutely continuous if f is continuous and
for t ∈ (qb, b], there exists a constant M > 0 such that
∞
∑
⏐f (tq j ) − f (tq j+1 )⏐ ≤ M.
⏐ ⏐

j=0
(n)
Introduce the space A Cq [0, b] = Dqk−1 f ∈ A Cq [0, b], k = 1, 2, . . . , n , where A Cq [0, b] is composed of
{ }

all functions which are q-absolutely continuous on [0, b].

The following lemma gives the relation between the Caputo type q-fractional derivative and the Riemann–
Liouville type q-fractional derivative which will be used in next section.

(n)
Lemma 2.2 ([5]). Let α > 0, n = ⌈α⌉ and t ∈ (0, b]. If f ∈ A Cq [0, b] and Dqn f ∈ C[0, b], then
⎛ ⎞
n−1
∑ Dqj f (0)
c α
Dq f (t) = Dqα ⎝f (t) − tj ⎠ . (2.12)
j=0
Γ q (j + 1)

3. The difference method for the Caputo type q-fractional differential equation

The initial value problem of the α-order (0 < α < 1) Caputo type q-fractional differential equation with
0 < q < 1 can be stated as follows.
c
Dqα x(t) = f (t, x(t)), t0 < t ≤ 1, x(t0 ) = x0 , (3.1)
n n
where t0 ∈ Tq and f : Tq × R → R is continuous.
 T. Zhang and Y. Tang / Applied Mathematics Letters 98 (2019) 292–299 295

The main task of this article is to develop the numerical method for the Caputo type q-fractional
differential equations. We first construct a difference formula to discretize the q-fractional derivative of
Caputo type.
According to (2.12), we have
x(0)
c α
Dq x(t) = Dqα x(t) − t−α (3.2)
Γq (1 − α)
Let 0 = t0 < t1 < · · · < tN = 1 be a nonuniform partition of [0, 1] with tk = q N −k ∈ Tq (1 ≤ k ≤ N ) and
step sizes ∆t1 = q N −1 , ∆tk = q N −k − q N −k+1 (2 ≤ k ≤ N ), where N ≥ 1 is a positive integer. At node tn ,
we have from (3.2), (2.10) and (2.8) (noting that when f (t, s) = (t − qs)(−α) x(s), f (qt, t) = 0) that
∫ tn
x(0) 1 x(0)
c α
Dq x(tn ) = Dq Iq1−α x(tn ) − t−α
n = Dq (tn − qs)(−α) x(s)dq s − t−α
Γq (1 − α) Γq (1 − α) 0 Γq (1 − α) n
∫ tn
1 x(0)
= Dq (tn − qs)(−α) x(s)dq s − t−α .
Γq (1 − α) 0 Γq (1 − α) n

According to (2.5) and (2.3), we further obtain

∫ tn
1 x(0)
c α
Dq x(tn ) = (tn − qs)(−α−1) x(s)dq s − t−α
Γq (−α) 0 Γq (1 − α) n
n ∫ tk
1 ∑ x(0)
= (tn − qs)(−α−1) x(s)dq s − t−α
n . (3.3)
Γq (−α) tk−1 Γ q (1 − α)
k=1

We need to further discretize the integral in (3.3). Introduce the piecewise linear interpolation of x(s):
s − tk−1 tk − s
L1,k (s) = x(tk ) + x(tk−1 ), s ∈ [tk−1 , tk ], k = 1, 2, . . . , N. (3.4)
∆tk ∆tk
Denote the interpolation error by R(tk−1 , tk , s) = x(s) − L1,k (s), k = 1, 2, . . . , N . In order to estimate
R(tk−1 , tk , s), we introduce the following lemma:

Lemma 3.1 (q-Rolle Lemma [26]). Let x(t) be a continuous function on [a, b], x(a) = x(b). For q̂ ∈ (0, 1),
∀ q ∈ (q̂, 1) ∪ (1, q̂ −1 ), then there exists ξ ∈ (a, b) such that Dq x(ξ) = 0.

By means of Lemma 3.1, we can establish the following theorem:

Theorem 3.1. Assume that Dq x(s) is continuous on [tk−1 , tk ] and Dq2 x(s) exists in (tk−1 , tk ). Then it holds
that
1
R(tk−1 , tk , s) = x(s) − L1,k (s) = D2 x(ξk )(s − tk−1 )(s − tk ), (3.5)
1+q q
where s ∈ [tk−1 , tk ], ξk ∈ (tk−1 , tk ), 1 ≤ k ≤ N .

Proof . Since R(tk−1 , tk , tk−1 ) = R(tk−1 , tk , tk ) = 0, we may assume that

R(tk−1 , tk , s) = Kq (s)(s − tk−1 )(s − tk ), s ∈ [tk−1 , tk ], (3.6)

where Kq (s) is an unknown function. For a fixed point t ∈ (tk−1 , tk ), let φq (s) = x(s) − L1,k (s) − Kq (t)(s −
tk−1 )(s − tk ). Since φq (tk−1 ) = φq (tk ) = φq (t) = 0, by using the q-Rolle Lemma twice, we know that there
exists ξk ∈ (tk−1 , tk ) such that

Dq2 φq (ξk ) = Dq2 x(ξk ) − (1 + q)Kq (t) = 0.

296 T. Zhang and Y. Tang / Applied Mathematics Letters 98 (2019) 292–299

So, we have
1
Kq (t) = D2 x(ξk ), ξk ∈ (tk−1 , tk ).
1+q q
Substituting Kq (t) into (3.6), we derive error formula (3.5). □
Now, on each interval [tk−1 , tk ], replacing x(s) by L1,k x(s), we obtain from (3.3)

∑ x(tk ) n ∫ tk
1 x(0)
c
Dqα x(tn ) = (tn − qs)(−α−1) (s − tk−1 )dq s − t−α
Γq (−α) ∆tk tk−1 Γq (1 − α) n
k=1
n ∫ tk
1 ∑ x(tk−1 )
+ (tn − qs)(−α−1) (tk − s)dq s + r(tn )
Γq (−α) ∆tk tk−1
k=1
n
1 ∑ 1 (α) (α) x(0)
= (b x(tk ) − bk,k x(tk−1 )) − t−α + r(tn ), (3.7)
Γq (−α) ∆tk k,k−1 Γq (1 − α) n
k=1

where the coefficient

∫ tk
(α)
bk,m = (tn − qs)(−α−1) (s − tm )dq s, k = 1, 2, . . . , n, m = k − 1, k,
tk−1

and the truncation error

n ∫ tk
1 ∑
r(tn ) = (tn − qs)(−α−1) R(tk−1 , tk , s)dq s. (3.8)
Γq (−α) tk−1
k=1

From (2.6) and (2.7), we have

∫ tk ∫ tk−1
(α) (−α−1)
bk,m = (tn − qs) (s − tm )dq s − (tn − qs)(−α−1) (s − tm )dq s
0 ( ∞ 0
∑
= (1 − q) tk q i (tn − q i+1 tk )(−α−1) (q i tk − tm )
i=0
∞
)
∑
i i+1 (−α−1) i
−tk−1 q (tn − q tk−1 ) (q tk−1 − tm ) . (3.9)
i=0

From (3.7) we obtain

c
Dqα x(tn ) = ∆α
q x(tn ) + r(tn ), (3.10)

where the difference formula:

n
1 ∑ 1 (α) (α) x(0)
∆α
q x(tn ) = (b x(tk ) − bk,k x(tk−1 )) − t−α . (3.11)
Γq (−α) ∆tk k,k−1 Γq (1 − α) n
k=1

Using difference formula (3.11), we present the difference method for solving problem (3.1) as follows.
{ α n
∆q x = f (tn , xn ), n = 1, 2, . . . , N,
(3.12)
x0 = x0 .

Lemma 3.2. Let 0 < α, q < 1. For two points s, t ∈ A and 0 ≤ s ≤ t, it holds that
⏐ ⏐ t−α−1
⏐(t − qs)(−α−1) ⏐ < . (3.13)
⏐ ⏐
(1 − q α )(1 − q 1−α )
 T. Zhang and Y. Tang / Applied Mathematics Letters 98 (2019) 292–299 297

Proof . We have from (2.1) that

∞ ⏐ ⏐ ∞ ⏐
⏐ t − q i+1 s ⏐ ⏐ ∏ ⏐ t − q i+1 s ⏐
⏐ ⏐ ⏐
⏐ = t−α−1 ⏐ t − qs ⏐
⏐ ⏐ ∏
⏐(t − qs)(−α−1) ⏐ = t−α−1
⏐
⏐ t − q i−α s ⏐ .
⏐ ⏐ ⏐ ⏐ ⏐
⏐ t − q i−α s ⏐ ⏐ t − q −α s ⏐
i=0 i=1

⏐ t − q i+1 s ⏐ i+1
⏐ ⏐ { ⏐ ⏐} ⏐ ⏐
⏐ t − qs ⏐ ⏐ 1−q ⏐ 1−q ⏐= 1−q
max ⏐⏐ ⏐ 1 − q −α ⏐ < 1 − q α , , i = 1, 2....
⏐ = max 1, ⏐ ⏐ max ⏐⏐
0≤s≤t t − q −α s ⏐ 0≤s≤t t − q i−α s ⏐ 1−q i−α

So, we have
∞ ⏐
⏐ t − q i+1 s ⏐ 2 3
⏐
−α−1 1 − q ⏐ = t−α−1 (1 − q) (1 − q ) (1 − q ) · · ·
⏐ ⏐ ∏
(−α−1) ⏐
⏐(t − qs) ⏐<t
⏐ ⏐
1 − q α i=1 t − q i−α s
⏐ ⏐ (1 − q α ) (1 − q 1−α ) (1 − q 2−α )
1 1 (1 − q) (1 − q 2 ) t−α−1
= t−α−1 α 1−α 2−α 3−α
··· < α
.
(1 − q ) (1 − q ) (1 − q ) (1 − q ) (1 − q )(1 − q 1−α )
The proof is completed. □

For the truncation error of difference formula (3.11), we have the following theorem.

Theorem 3.2. Let 0 < α, q < 1. Assume that x(t) is continuous on [0, 1] and M1 = sup ⏐Dq2 x(t)⏐. Then,
⏐ ⏐
0≤t≤1
for any fixed δ (0 < δ < 1) and at mesh point tn with n ≤ (1 − δ)N , the truncation error r(tn ) satisfies the
following estimation:
M1 (1 − q)2 q (2−α)δN
|r(tn )| < . (3.14)
4(1 − q α )(1 − q 1−α ) |Γq (−α)|

Proof . From (3.8) and (3.5) we have

⏐ n ∫ tk
⏐
⏐ 1 ∑
(−α−1)
⏐
|r(tn )| = ⏐ −(tn − qs) R(tk−1 , tk , s)dq s⏐
⏐ ⏐
⏐ Γq (−α)
k=1 tk−1
⏐
⏐ ⏐
n ∫ tk
⏐ M1 ∑ ⏐
≤⏐ (tn − qs)(−α−1) (s − tk−1 )(tk − s)dq s⏐
⏐ ⏐
⏐ (1 + q)Γq (−α)
k=1 tk−1
⏐
n ∫ tk n ∫ tk
M1 ∑ M1 ∑
< ∆t2k |(tn − qs)(−α−1) |dq s < ∆t2k |(tn − qs)(−α−1) |dq s.
4|Γq (−α)| tk−1 4|Γq (−α)| tk−1
k=1 k=1

According to Lemma 3.2, we have

∫ tn
M1 −α−1 2
|r(tn )| < t max ∆tk dq s
4(1 − q α )(1 − q 1−α ) |Γq (−α)| n 1≤k≤n 0
M1
= t−α max ∆t2
4(1 − q )(1 − q 1−α ) |Γq (−α)| n 1≤k≤n k
α

M1
= q −α(N −n) max ((1 − q)q N −k )2
4(1 − q α )(1 − q 1−α ) |Γq (−α)| 1≤k≤n
2
M1 (1 − q)
< q (2−α)(N −n)
4(1 − q α )(1 − q 1−α ) |Γq (−α)|
M1 (1 − q)2 q (2−α)δN
< .
4(1 − q α )(1 − q 1−α ) |Γq (−α)|
The proof is completed. □

From the truncation error estimation (3.14), we see that for any fixed node tn ∈ (0, 1), it holds that
r(tn ) → 0 as N → ∞.
298 T. Zhang and Y. Tang / Applied Mathematics Letters 98 (2019) 292–299

Table 1
Numerical solutions and errors (N = 20, q = 3/5, α = 1/2).
tn x(tn ) xn |x(tn ) − xn |
0 0 0 0
(3/5)19 3.89570E−7 4.34055E−7 4.44849E−8
(3/5)17 1.80357E−6 1.82651E−6 2.29471E−8
(3/5)15 8.34985E−6 8.47005E−6 1.20204E−7
(3/5)13 3.86567E−5 3.87139E−5 5.72027E−8
(3/5)11 1.78966E−4 1.78998E−4 3.23198E−8
(3/5)9 8.28547E−4 8.28566E−4 1.90138E−8
(3/5)7 3.83587E−3 3.83588E−3 1.13302E−8
(3/5)5 1.77586E−2 1.77586E−2 6.78151E−9
(3/5)3 8.22159E−2 8.22159E−2 4.06534E−9
3/5 3.80629E−1 3.80629E−1 1.22486E−7
1 8.18984E−1 8.18984E−1 5.42886E−8

Table 2
Numerical solutions and errors (N = 10, q = 1/2, α = 1/3).
tn x(tn ) xn x̄n |x(tn ) − xn | |x(tn ) − x̄n |
0 1.00000E−4 1.00000E−4 1.00000E−4 0 0
(1/2)9 8.55796E−4 6.20246E−4 1.07045E−4 2.35550E−4 7.48751E−4
(1/2)8 1.79670E−3 1.64993E−3 1.02899E−4 1.46778E−4 1.69380E−3
(1/2)7 3.90897E−3 3.74731E−3 1.01251E−4 1.61662E−4 3.80772E−3
(1/2)6 8.65085E−3 7.96510E−3 1.00549E−4 6.85748E−4 8.55030E−3
(1/2)5 1.92960E−2 1.93244E−2 1.00243E−4 2.83694E−5 1.19196E−2
(1/2)4 4.31936E−2 4.30172E−2 1.00108E−4 1.76373E−4 4.30935E−2
(1/2)3 9.68418E−2 9.66938E−2 1.00048E−4 1.48068E−4 9.67418E−2
(1/2)2 2.17278E−1 2.17122E−1 1.00021E−4 1.56392E−4 2.17178E−1
1/2 4.87648E−1 4.87492E−1 1.00010E−4 1.55986E−4 4.87548E−1
1 1.09461 1.09445 1.00004E−4 1.56535E−4 1.09451

4. Numerical experiment

In this section, we give two examples to show the effectiveness of the difference method. Using difference
scheme (3.12) and Mathematica 11.3 to carry out the computation, we obtain the numerical results shown
in Tables 1 and 2 respectively.

Example 1. Consider the Caputo type q-fractional initial value problem

1
c
Dq2 x(t) = t, 0 < t ≤ 1, x(0) = 0. (4.1)

The exact solution of Eq. (4.1) is x(t) = Γq (2)t3/2 /Γq (5/2).

Example 2. Consider the Caputo type q-fractional initial value problem

1
c
Dq3 x(t) = t−3/2 x2 (t), 0 < t ≤ 1, x(0) = 0.0001. (4.2)

The exact solution of Eq. (4.2) is x(t) = Γq (13/6)t7/6 /Γq (11/6) + 0.0001.
In Table 2, we also list the numerical solution x̄n obtained by using the iteration method [20]. We can see
that the difference method has a high precision and it remains valid as t closes to 1.

Acknowledgment

This work was supported by the State Key Laboratory of Synthetical Automation for Process Industries,
China Fundamental Research Funds [grant number 2013ZCX02].
 T. Zhang and Y. Tang / Applied Mathematics Letters 98 (2019) 292–299 299

References

[1] F.H. Jackson, On q-functions and a certain difference operator, Trans. R. Soc. Edinb. 46 (1908) 64–72.
[2] W.A. Al-Salam, Some fractional q-integrals and q-derivatives, Proc. Edinb. Math. Soc. (2). 15 (1966) 135–140.
[3] R.P. Agarwal, Certain fractional q-integrals and q-derivatives, Proc. Camb. Phil. Soc. 66 (1969) 365–370.
[4] W. Hahn, Beiträge zur theorie der Heineschen Reihen, Math. Nachr. 2 (1949) 340–379.
[5] M.H. Annaby, Z.S. Mansour, q-Fractional Calculus and Equations, Springer Heidelberg, New York, 2012.
[6] G.E. Andrews, R. Askey, R. Roy, Special Functions, Cambridge University Press, 1999.
[7] V. Kac, P. Cheung, Quantum Calculus, Springer-Verlag, New York, 2002.
[8] T. Abdeljawad, J. Alzabut, On Riemann–Liouville fractional q-difference equations and their application to retarded
logistic type model, Math. Methods Appl. Sci. 41 (2018) 8953–8962.
[9] B. Abdalla, On the oscillation of q-fractional difference equations, Adv. Differential Equations 1 (2017) 254–264.
[10] H. Aktuğlu, M.A. Özarslan, On the solvability of Caputo q-fractional boundary value problem involving p-Laplacian
operator, Abstr. Appl. Anal. 2013 (2013) 1–8.
[11] R. Almeida, N. Martins, Existence results for fractional q-difference equations of order α ∈ ]2,3[ with three-point
boundary conditions, Commun. Nonlinear Sci. Numer. Simul. 19 (2014) 1675–1685.
[12] H.Q. Chen, S.G. Kang, L.L. Kong, et al., Existence of three positive solutions for a class of boundary value problems
of Caputo fractional q-difference equation, Discret. Dyn. Nat. Soc. 2 (2018) 1–9.
[13] B. Ahmad, S.K. Ntouyas, I.K. Purnaras, Existence results for nonlocal boundary value problems of nonlinear fractional
q-difference equations, Adv. Differential Equations 2012 (2012) 140–154.
[14] F. Jarad, T. Abdeljawad, D. Baleanu, Stability of q-fractional non-autonomous systems, Nonlinear Anal. RWA 14 (2013)
780–784.
[15] G.T. Wang, W. Sudsutad, L.H. Zhang, et al., Monotone iterative technique for a nonlinear fractional q-difference equation
of Caputo type, Adv. Differential Equations 1 (2016) 211–222.
[16] T. Abdeljawad, B. Benli, D. Baleanu, A generalized q-Mittag-Leffler function by q-Caputo fractional linear equations,
Abstr. Appl. Anal. 2012 (2012) 1–11.
[17] T. Abdeljawad, D. Baleanu, Caputo q-fractional initial value problems and a q-analogue Mittag-Leffler function,
Commun. Nonlinear Sci. Numer. Simul. 16 (2011) 4682–4688.
[18] S. Salahshour, A. Ahmadian, C.S. Chan, Successive approximation method for Caputo q-fractional IVPs, Commun.
Nonlinear Sci. Numer. Simul. 24 (2015) 153–158.
[19] G.C. Wu, D. Baleanu, New applications of the variational iteration method - from differential equations to q-fractional
difference equations, Adv. Differential Equations 1 (2013) 21–37.
[20] Y.C. Tang, T. Zhang, A remark on the q-fractional differential equations, Appl. Math. Comput. 350 (2019) 198–208.
[21] R.A.C. Ferreira, Nontrivial solutions for fractional q-difference boundary value problems, Electron. J. Qual. Theory
Differ. Equ. 70 (2010) 1–10.
[22] R. Askey, The q-gamma and q-beta functions, Appl. Anan. 8 (1978) 125–141.
[23] F.M. Atici, P.W. Eloe, Fractional q-calculus on a time scale, J. Math. Phys. 14 (2007) 341–352.
[24] F.H. Jackson, On q-definite integrals, Quart. J. Pure Appl. Math. 41 (1910) 193–203.
[25] P.M. Rajković, S.D. Marinković, M.S. Stanković, On q-analogues of Caputo derivative and Mittag-Leffler function, Fract.
Calc. Appl. Anal. 10 (2007) 359–373.
[26] P.M. Rajković, M.S. Stanković, S.D. Marinković, Mean value theorems in q-calculus, Math. Vesnic θ 54 (2002) 171–178.