fractal and fractional

Article
Numerical Solution of Variable-Order Fractional Differential
Equations Using Bernoulli Polynomials
Somayeh Nemati 1,† , Pedro M. Lima 2,† and Delfim F. M. Torres 3, *,†

1 Department of Applied Mathematics, Faculty of Mathematical Sciences, University of Mazandaran,

Babolsar P.O. Box 47416-95447, Iran; [email protected]
2 Centro de Matemática Computacional e Estocástica, Instituto Superior Técnico, Universidade de Lisboa,
Av. Rovisco Pais, 1049-001 Lisboa, Portugal; [email protected]
3 Center for Research and Development in Mathematics and Applications (CIDMA), Department of
Mathematics, University of Aveiro, 3810-193 Aveiro, Portugal
* Correspondence: [email protected]
† These authors contributed equally to this work.

Abstract: We introduce a new numerical method, based on Bernoulli polynomials, for solving
multiterm variable-order fractional differential equations. The variable-order fractional derivative
was considered in the Caputo sense, while the Riemann–Liouville integral operator was used to give
approximations for the unknown function and its variable-order derivatives. An operational matrix of
variable-order fractional integration was introduced for the Bernoulli functions. By assuming that the
solution of the problem is sufficiently smooth, we approximated a given order of its derivative using
Bernoulli polynomials. Then, we used the introduced operational matrix to find some approximations
for the unknown function and its derivatives. Using these approximations and some collocation




points, the problem was reduced to the solution of a system of nonlinear algebraic equations. An

 error estimate is given for the approximate solution obtained by the proposed method. Finally,
Citation: Nemati, S.; Lima, P.M.; five illustrative examples were considered to demonstrate the applicability and high accuracy of
Torres, D.F.M. Numerical Solution of the proposed technique, comparing our results with the ones obtained by existing methods in the
Variable-Order Fractional Differential literature and making clear the novelty of the work. The numerical results showed that the new
Equations Using Bernoulli method is efficient, giving high-accuracy approximate solutions even with a small number of basis
Polynomials. Fractal Fract. 2021, 5, functions and when the solution to the problem is not infinitely differentiable, providing better results
219. https://doi.org/10.3390/ and a smaller number of basis functions when compared to state-of-the-art methods.
fractalfract5040219

Keywords: fractional differential equations; numerical methods; variable-order fractional calculus;

Academic Editor: Paul Eloe
operational matrix of variable-order fractional integration; Bernoulli polynomials

Received: 31 October 2021

MSC: 34A08; 65L60
Accepted: 12 November 2021
Published: 14 November 2021

Publisher’s Note: MDPI stays neutral

1. Introduction
with regard to jurisdictional claims in
published maps and institutional affil- In the last few decades, fractional calculus has attracted the attention of many scien-
iations. tists in different fields such as mathematics, physics, chemistry, and engineering. Due to
the fact that fractional operators consider the evolution of the system, by taking the global
correlation, and not only local characteristics, some physical phenomena are better de-
scribed by fractional derivatives [1]. There are many definitions of fractional differentiation
Copyright: © 2021 by the authors.
and integration in the literature (for details, see, e.g., [2–5]). However, the most commonly
Licensee MDPI, Basel, Switzerland.
used are the definitions of the Caputo derivative and Riemann–Liouville fractional integral
This article is an open access article
operators. For new fractional derivatives with nonlocal and nonsingular kernels, with
distributed under the terms and applications in rheological models, we refer the reader to [6].
conditions of the Creative Commons A recent generalization of the theory of fractional calculus is to allow the fractional
Attribution (CC BY) license (https:// order of the derivatives to be dependent on time, i.e., to be nonconstant or of variable order.
creativecommons.org/licenses/by/ In [7], the authors investigated operators when the order of the fractional derivative is
4.0/). variable with time [8,9].

Fractal Fract. 2021, 5, 219. https://doi.org/10.3390/fractalfract5040219 https://www.mdpi.com/journal/fractalfract

Fractal Fract. 2021, 5, 219 2 of 16

The nonlocal properties of systems are more visible with variable-order fractional
calculus, and many real-world phenomena in physics, mechanics, control, and signal
processing have been described by this approach [10–13]. In particular, several applications
of variable-order fractional calculus are found in engineering mechanics; see [14] for
an application of variable-order fractional operators to model the microscopic structure
of a material, [15] for an application of the Riesz–Caputo fractional derivative of space-
dependent order in continuum elasticity, [16,17] for the nonlinear viscoelastic behavior of
fractional systems with variable time-dependent fractional order, and [18] for the use of
variable-order fractional calculus in the static response of nonlocal beams having either a
porous or a functionally graded core.
Obtaining analytic solutions for such fractional differential equations (FDEs) is, how-
ever, very difficult. Therefore, in most cases, the exact solution is not known, and one
needs to seek a numerical approximation. Therefore, many researchers have introduced
and developed numerical methods in order to obtain approximated solutions for this class
of equations. For example, in [19,20], Legendre polynomials were used to construct a
numerical solution for a class of multiterm variable-order FDEs, while [21] used Legendre
wavelets and operational matrices; References [22,23] used Chebyshev polynomials, re-
spectively of the second and fourth kinds; Reference [24] used Bernstein polynomials. The
book [25] showed the usefulness of numerical methods for approximating variable-order
fractional operators in the framework of the calculus of variations; the paper [26] adopted
Coimbra’s variable-order time-fractional operator, discussing the stability, convergence and
solvability of a numerical scheme based on Fourier analysis. Reference [27] implemented a
numerical method for solving a circulant Halvorsen system described by Caputo fractional
variable-order derivatives. For multivariable-order differential equations with nonlocal
and nonsingular kernels, we refer to [28], where a collocation method was developed based
on Chebyshev polynomials of the fifth kind.
Here, we considered the following general form of a multiterm variable-order FDE:

C α(t)
 
α (t) α (t) α (t)
0 Dt y(t) = F t, y(t), 0C Dt 1 y(t), 0C Dt 2 y(t), . . . , 0C Dt k y(t) , 0 < t ≤ 1, (1)

with initial conditions

y(i) (0) = y0i , i = 0, 1, . . . , n − 1, (2)
where n is the smallest positive integer number such that for all t ∈ [0, 1], one has 0 <
α(t) α (t) α (t)
α(t) ≤ n, 0 < α1 (t) < α2 (t) < . . . < αk (t) < α(t), and 0C Dt , 0C Dt 1 ,. . . , 0C Dt k are the
(left) fractional derivatives of variable-order defined in the Caputo sense. Problems of
the form (1) and (2) have a practical impact. Indeed, specific applications are found in
noise reduction and signal processing [29,30], the processing of geographical data [31], and
signature verification [32].
In recent years, Bernoulli polynomials have been shown to be a powerful mathematical
tool in dealing with various problems of a dynamical nature, e.g., for solving numerically
high-order Fredholm integrodifferential equations [33], pantograph equations [34], partial
differential equations [35], linear Volterra and nonlinear Volterra–Fredholm–Hammerstein
integral equations [36], as well as optimal control problems [37]. Here, we employ a
spectral method based on Bernoulli polynomials in order to obtain numerical solutions
to the problem (1) and (2). Our method consists of reducing the problem to a system of
nonlinear algebraic equations. To do this, we introduced an accurate operational matrix of
variable-order fractional integration for the Bernoulli polynomials’ basis vector. To the best
of our knowledge, this is the first time in the literature that such a method for solving a
general class of multiterm variable-order FDEs based on the Riemann–Liouville fractional
integral of the basis vector has been introduced.
The rest of this paper is organized as follows. In Section 2, some preliminaries
of variable-order fractional calculus are reviewed and some properties of the Bernoulli
polynomials are recalled. Section 3 is devoted to introducing the operational matrix of
variable-order fractional integration for Bernoulli polynomials. In Section 4, we present
Fractal Fract. 2021, 5, 219 3 of 16

a new numerical method for solving the problem (1) and (2) by using the operational
matrix technique and collocation points. Section 5 is concerned with presenting an error
estimate for the numerical solution obtained by this new scheme. In Section 6, several
multiterm variable-order FDEs are considered and solved, using the introduced method.
Finally, concluding remarks are given in Section 7, where some possible future directions
of research are also pointed out.

2. Preliminaries
In this section, a brief review of the necessary definitions and properties of the variable-
order fractional calculus is presented. Furthermore, Bernoulli polynomials, and some of
their properties, are recalled.

2.1. Some Preliminaries of Variable-Order Fractional Calculus

We followed the notations of [1].

Definition 1 (See, e.g., [1]). The left Riemann–Liouville fractional integral of order α(t) is
defined by:
Z t
α(t) 1
0 It y(t) = (t − s)α(t)−1 y(s)ds, t > 0,
Γ(α(t)) 0
where Γ(·) is the Euler gamma function.

Lemma 1 (See Chapter 1, Lemma 8, of [1]). Let y be the power function y(t) = tν . Then, for
ν > −1, we have:
α(t) Γ ( ν + 1)
0 It y(t) = tν+α(t) , α(t) ≥ 0.
Γ(ν + 1 + α(t))

Definition 2 (See, e.g., [1]). The left Caputo fractional derivative of order α(t) is defined by:
Z t
C α(t) 1
0 Dt y(t) = (t − s)n−α(t)−1 y(n) (s)ds, n − 1 < α(t) < n,
Γ(n − α(t)) 0
C α(t)
0 Dt y(t) = y ( n ) ( t ), α(t) = n.

For 0 ≤ α(t) ≤ n, n ∈ N, and γ > 0, two useful properties of the Caputo derivative
and Riemann–Liouville integral are:

dγe−1
ti
∑
γ C γ
0 It (0 D t y ( t )) = y(t) − y (i ) (0) , t > 0, (3)
i =0
i!

n −1
n−α(t) (n) α(t) ti −α(t)
0 It (y (t)) = 0C D t y(t) − ∑ y (i ) (0)
Γ(i + 1 − α(t))
, t > 0, (4)
i =dα(t)e

where d·e is the ceiling function.

2.2. Bernoulli Polynomials

Bernoulli polynomials build a family of independent polynomials that form a complete
basis for the space L2 [0, 1], which is the space of all square integrable functions on the
interval [0, 1]. The Bernoulli polynomial of degree m, β m (t), is defined as follows [38]:
m 
m
β m (t) = ∑ bm − i t i , (5)
i =0
i

where bk , k = 0, 1, . . . , m, are the Bernoulli numbers that appear in the series expansion of
trigonometric functions [39] and can be defined by the following identity:
Fractal Fract. 2021, 5, 219 4 of 16

∞
t ti
= ∑ bi .
et − 1 i=0 i!

The first four Bernoulli polynomials are:

β 0 (t) = 1,
1
β 1 (t) = t − ,
2
1
β 2 ( t ) = t2 − t + ,
6
3 1
β 3 (t) = t3 − t2 + t.
2 2
The Bernoulli polynomials satisfy the following property [39]:
Z 1
i!j!
β i (t) β j (t)dt = (−1)i−1 b , i, j ≥ 1. (6)
0 (i + j ) ! i + j

Any arbitrary function y ∈ L2 [0, 1] can be approximated using the Bernoulli polyno-
mials as
M
y(t) ' ∑ a m β m ( t ) = A T B ( t ), (7)
m =0

where
B(t) = [ β 0 (t), β 1 (t), . . . , β M (t)] T (8)
and
A = [ a0 , a1 , . . . , a M ] T .
The coefficient vector A in (7) is calculated by the following formula (see [37]):

A = D −1 hy(t), B(t)i, (9)

where h·, ·i is the inner product, defined for two arbitrary functions f , g ∈ L2 [0, 1] as
Z 1
h f (t), g(t)i = f (t) g(t)dt
0

and D = h B(t), B(t)i, which is calculated using (6).

3. Operational Matrix of Variable-Order Fractional Integration

The aim of this section is to introduce an accurate operational matrix of variable-order
fractional integration for Bernoulli functions. To do this, we display the Bernoulli basis
vector B(t), given by (8), in terms of the Taylor basis functions, as follows:

B(t) = QT(t), (10)

where T is the Taylor basis vector

T(t) = [1, t, t2 , . . . , t M ] T ,

and Q is the change-of-basis matrix, which is obtained using (5) as

Fractal Fract. 2021, 5, 219 5 of 16

 
1 0 0 0 0 ... 0
 −1 1 0 0 0 ... 0 
 2 
 1 −1 1 0 0 ... 0 
 6 
Q=
 0 1
− 23 1 0 ... 0
.
 . 2 
 . .. .. .. .. .. 

 . . . . . . 
bM (M
1 ) b M −1 (M
2 ) b M −2 (M
3 ) b M −3 (M
4 ) b M −4 ... 1

It is obvious that the matrix Q is a nonsingular matrix. Therefore, we can write:

T(t) = Q−1 B(t). (11)

Taking (10) into account and by applying the left Riemann–Liouville fractional integral
operator of order α(t) to the vector B(t), we obtain

α(t) α(t) α(t) α(t)

0 It B(t) = 0 It ( QT(t)) = Q(0 It T(t)) = QSt T(t), (12)

α(t)
where St is a diagonal matrix, which is obtained using Lemma 1 as follows:

1
tα(t) ···
 
Γ(1+α(t))
0 0 0 0
1

 0 Γ(2+α(t))
tα(t) 0 0 ··· 0 

2
tα(t)
 
St
α(t)
=
 0 0 Γ(3+α(t))
0 ··· 0 .

 .. .. .. .. .. 
. . . . .
 
 
Γ ( M +1)
0 0 0 0 ··· Γ( M +1+α(t))
tα(t)

Finally, by using (11) into (12), we have:

α(t) α(t) α(t)

0 It B(t) = QSt Q−1 B(t) = Pt B ( t ), (13)

α(t) α(t)
where Pt = QSt Q−1 is a matrix of dimension ( M + 1) × ( M + 1), which we call
the operational matrix of variable-order fractional integration of order α(t) of Bernoulli
α(t)
functions. Since Q and Q−1 are lower triangular matrices and St is a diagonal matrix, it
α(t)
is obvious that Pt is also a lower triangular matrix. For example, with M = 2, one has:

p1,1 0 0
 
1
Pt
α(t)  p2,1
= Γ(α(t)+2)
tα(t) 0 ,

 
2 1 2
p3,1 Γ(α(t)+3)
− Γ(α(t)+2)
tα(t) Γ(α(t)+3)
tα(t)

where
1
p1,1 = tα(t) ,
Γ ( α ( t ) + 1)
 
1 1
p2,1 = − tα(t) ,
2Γ(α(t) + 2) 2Γ(α(t) + 1)
 
1 1 2
p3,1 = − + tα(t) .
6Γ(α(t) + 1) 2Γ(α(t) + 2) 3Γ(α(t) + 3)

4. Numerical Method
This section is devoted to presenting a new numerical method for solving the multi-
term variable-order FDE (1) with initial conditions (2). To this aim, we set:

n := max {dα(t)e}.
0< t ≤1
Fractal Fract. 2021, 5, 219 6 of 16

Then, by assuming y ∈ C n [0, 1], we considered an approximation of the n-th order

derivative of the unknown function y using the Bernoulli functions as follows:

y ( n ) ( t ) = A T B ( t ), (14)

where A is an ( M + 1) × 1 vector with unknown elements and B(t) is the Bernoulli basis
vector given by (8). Taking into account the initial conditions (2) and using (3), (13) and
(14), we obtain:
n −1
ti
y(t) = 0 Itn (y(n) (t)) + ∑ y(i) (0) i!
i =0
n −1
ti
= A T (0 Itn B(t)) + ∑ y0i
i!
(15)
i =0
n −1
ti
= A T Ptn B(t) + ∑ y0i i! .
i =0

In a similar way, using (4), (13), and (14), we obtain

n −1
C α(t) n−α(t) ti −α(t)
0 Dt y(t) = A T Pt B(t) + ∑ y0i
Γ(i + 1 − α(t))
(16)
i =dα(t)e

and
C α j (t) n−α j (t)
n −1
ti −α j (t)
0 Dt y(t) = A T Pt B(t) + ∑ y0i
Γ(i + 1 − α j (t))
, j = 1, . . . , k. (17)
i =dα j (t)e

By substituting the approximations given in (15)–(17) into Equation (1), we have:


n −1 n −1
ti n − α1 ( t ) t i − α1 ( t )
F t, A T Ptn B(t) + ∑ y0i i! , AT Pt B(t) + ∑ y0i
Γ(i + 1 − α1 (t))
,
i =0 i =dα1 (t)e

n −1
n−αk (t) ti −αk (t)
. . . , A T Pt B(t) + ∑ y0i
Γ(i + 1 − αk (t))

i =dαk (t)e
n −1
n−α(t) ti −α(t)
= A T Pt B(t) + ∑ y0i
Γ(i + 1 − α(t))
. (18)
i =dα(t)e

j +1
By using (18) at M + 1 collocation points, which are chosen as t j = M+2 , with
j = 0, 1, . . . , M, we obtain the following system of nonlinear algebraic equations:

n−α(t j )
n −1 t j i −α(t j )
A T Pt j B(t j ) + ∑ y0i
Γ(i + 1 − α(t j ))
i =dα(t j )e
n −1 t j i T n − α1 ( t j )
= F t j , A T Ptnj B(t j ) + ∑ y0i
i!
, A Pt j B(t j )
i =0
(19)
n −1 t j i − α1 ( t j ) n−αk (t j )
+ ∑ y0i , . . . , A T Pt j B(t j )
i =dα1 (t j )e
Γ(i + 1 − α1 (t j ))

n −1 i −αk (t j )
t j
+ ∑ y0i .
i =dα (t )e
Γ ( i + 1 − α k ( t j ))
k j
Fractal Fract. 2021, 5, 219 7 of 16

System (19) includes M + 1 nonlinear algebraic equations in terms of the unknown

parameters of vector A. After solving this system, an approximation of the solution of the
problem (1) and (2) is given by (15).

5. Error Estimate
The purpose of this section is to obtain an estimate of the error norm for the approxi-
mate solution obtained by the proposed method in Section 4. We assume that f (t) = y(n) (t)
is a sufficiently smooth function on [0, 1] and q M (t) is the interpolating polynomial to f at
points ts , where ts , s = 0, 1, . . . , M, are the roots of the ( M + 1)-degree shifted Chebyshev
polynomial in [0, 1]. Then, according to the interpolation error, we have:

M
f ( M +1) ( ξ )
f (t) − q M (t) =
( M + 1) ! ∏ ( t − t s ), ξ ∈ (0, 1).
s =0

Therefore,
f ( M +1) ( t ) M
| f (t) − q M (t)| ≤ max
t∈(0,1) ( M + 1) ! ∏ | t − t s |. (20)
s =0

We assume that there is a real number κ such that:

max f ( M+1) (t) ≤ κ. (21)

t∈(0,1)

By using (21) in (20) and taking into consideration the estimates for Chebyshev inter-
polation nodes [40], we obtain:
κ
| f (t) − q M (t)| ≤ . (22)
22M+1 ( M + 1)!

As a consequence of (22), we obtain the following result.

Theorem 1. Let f M (t) = A T B(t) be the Bernoulli functions’ expansion of a sufficiently smooth
function f defined on [0, 1], where A and B(t) are given, respectively, by (8) and (9). Then, there
exists a real number κ such that:
κ
k f (t) − f M (t)k ≤ . (23)
22M+1 ( M + 1)!

Proof. Let Π M be the space of all polynomials of degree ≤ M on t ∈ [0, 1]. By definition,
f M is the best approximation of f in Π M . Therefore, we have:

k f (t) − f M (t)k2 ≤ k f (t) − g(t)k2 ,

where g is any arbitrary polynomial in Π M . Therefore, we can write:

Z 1 Z 1
k f (t) − f M (t)k22 = | f (t) − f M (t)|2 dt ≤ | f (t) − q M (t)|2 dt, (24)
0 0

where q M is the interpolating polynomial of f , as discussed before. As a result, taking into

consideration (22) in (24), we easily obtain (23).

With the help of Theorem 1 and by assuming f (t) = y(n) (t), we obtain the following
result.
Fractal Fract. 2021, 5, 219 8 of 16

Theorem 2. Let the exact solution y of the problem (1) and (2) be a real-valued sufficiently smooth
function and y M be the approximate solution of this problem obtained by the method proposed in
Section 4. Then, we have:
κ
ky(t) − y M (t)k2 ≤ p ,
22M+1 ( M + 1)!(n − 1)! 2n(2n − 1)

where y is the exact solution and κ = maxt∈(0,1) y(nM+n) (t) , n = max0<t≤1 {dα(t)e}.

Proof. Let X be the space of all real-valued functions defined on [0, 1] and 0 Itn : X → X be
the Riemann–Liouville integral operator. We use the following definition for the norm of
the operator 0 Itn :
k0 Itn k2 = sup k0 Itn gk2 .
k g k2 =1

To continue the proof, first, we introduce an upper bound for k0 Itn k2 . To this end, using
the definition of the left Riemann–Liouville integral operator and Schwarz’s inequality,
we obtain:
Z t 2
1
k0 Itn gk22 = (t − s)n−1 g(s)ds
( n − 1) ! 0 2
Z t 2 Z 1 Z t 2
1 1
= (t − s)n−1 g(s)ds = (t − s)n−1 g(s)ds dt
[(n − 1)!]2 0 2 [(n − 1)!]2 0 0
Z 1 Z t 1
Z 
1
≤ (t − s)2n−2 ds | g(s)|2 ds dt
[(n − 1)!]2 0 0 0
1
= .
[(n − 1)!]2 2n(2n − 1)

Therefore, we have:

1
k0 Itn k2 ≤ p . (25)
(n − 1)! 2n(2n − 1)

On the other hand, from Theorem 1, we have the following error bound:
κ
y(n) ( t ) − A T B ( t ) ≤ . (26)
2 22M+1 ( M + 1)!

Finally, using (3), (15), (25) and (26), we obtain that:

!
n −1 i n −1 i
t t
ky(t) − y M (t)k2 = n (n)
0 It ( y (t)) + ∑ y0i
i!
− n T
0 It ( A B ( t )) + ∑ y0i
i!
i =0 i =0 2
n (n)
= 0 It ( y (t) − A T B(t))
2

≤ k0 Itn k2 y (n)
(t) − A B(t)T
2
κ
≤ p ,
22M+1 ( M + 1)!(n − 1)! 2n(2n − 1)

which completes the proof.

6. Illustrative Examples
In this section, we apply our method to some variable-order FDEs and, moreover, to
one variable-order fractional pantograph differential equation (Example 4), comparing the
results with the ones obtained by existing methods in the literature. We implemented our
method and performed our numerical simulations with Mathematica 12.
Fractal Fract. 2021, 5, 219 9 of 16

Example 1. In our first example, we considered the following multiterm variable-order FDE [19,22,
23]:
t t t
C 2t 1
2C 3
1C
4
1
4C 5
1
0 D t y(t) + t 0 D t y(t) + t 0 D t y(t) + t 0 D t y(t) + t y(t)
3 5 = g ( t ), 0 < t ≤ 1, (27)

where
t t t
t2−2t t 2− 3 t 2− 4 t 2− 5 t2
 
1 1 1 1
g(t) = − − t2 − t 3 − t 4 + t 5 2 −
Γ(3 − 2t) Γ(3 − 3t ) Γ(3 − 4t ) Γ(3 − 5t ) 2

2
with initial conditions y(0) = 2 and y0 (0) = 0. The exact solution to this problem is y(t) = 2 − t2 .
As can be seen, we have α(t) = 2t. Therefore, to implement the proposed method, we introduce:

n = max {d2te} = 2.
0< t ≤1

We set M = 1 and suppose

y00 (t) = A T B(t),
where  T
T 1
A = [ a0 , a1 ] and B(t) = 1, t − .
2
The operational matrices of variable-order fractional integration are given as follows:

2− 3t
   
t2−2t t
0 2− 3t
0
Γ(3−2t) Γ(3− 3t )
Pt2−2t =  , Pt = ,
 
t2−2t t2−2t t2−2t 2− 3t 2− 3t 2− 3t
2Γ(4−2t)
− 2Γ(3−2t) Γ(4−2t)
t
− t t
2Γ(4− 3t ) 2Γ(3− 3t ) Γ(4− 3t )

2− t 2− 5t
   
t 4 t
2− 4t
0 2− 5t
0
Γ( 3− 4t ) Γ(3− 5t )
Pt = , Pt = .
   
2− t 2− t 2− t 2− 5t 2− t 2− 5t
t 4 t 4 t 4 t t 5 t
− −
2Γ(4− 4t ) 2Γ(3− 4t ) Γ 4− 4t
( ) 2Γ(4− 5t ) 2Γ(3− 5t ) Γ(4− 5t )

Furthermore, we have: " #

t2
0
Pt2 = 2
2 t2
.
− t6 6

Now, using the initial conditions and the aforementioned operational matrices, we obtain the
following approximations for y(t) and its variable-order derivatives:

C 2t
1
ti−2t
y(t) = A T Pt2 B(t) + 2, 0 D t y(t) = A T Pt2−2t B(t) + ∑ y0i
Γ(i − 2t + 1)
,
i =d2te

t
t
2− 3t
1
ti − 3
C 3
0 D t y(t) = A T Pt B(t) + ∑ y0i
Γ(i − 3t + 1)
,
i =d 3t e
t
t
2− t
1
ti − 4
C 4
0 D t y(t) = A T Pt 4 B(t) + ∑ y0i
Γ(i − 4t + 1)
,
i =d 4t e
t
t
2− 5t
1
ti − 5
C 5
0 D t y(t) = A T Pt B(t) + ∑ y0i
Γ(i − 5t + 1)
.
i =d 5t e

By substituting these approximations into (27), collocating the resulting equation at t0 = 13 ,

t1 = 32 , and finally, solving the obtained system, we obtain:

a0 = −1, a1 = 0.
Fractal Fract. 2021, 5, 219 10 of 16

Therefore, we have:
" #
t2 t2

2 0 1
y(t) = [−1 0] 2 t2 1 +2 = 2− ,
− t6 t− 2 2
6

which is the exact solution. In this case, since the exact solution is a second-order polynomial, we
can obtain it by applying the numerical method with just two basis functions.

Example 2. In our second example, we considered the following nonlinear variable-order FDE
borrowed from [41]:

C α(t)
0 Dt y(t) + sin(t)y2 (t) = g ( t ), 0 < t ≤ 1, 0 < α(t) ≤ 1,
y(0) = 0,

where
Γ( 92 ) 7
g(t) =
t 2 −α(t) + sin(t)t7 .
Γ 9
2 − α(t)
7
The exact solution of this problem is y(t) = t 2 . By considering α(t) = 1 − 0.5 exp(−t), we
solved this problem with different values of M. The numerical results are displayed in Figure 1 and
Table 1. In Figure 1, the approximate solutions obtained with M = 1, 2, 3, together with the exact
solution of this problem, are plotted. Furthermore, by considering M = 2, 6, 10, the absolute errors
at some selected points are reported in Table 1. From these results, the convergence of the numerical
solutions to the exact one can be easily seen.

1.0
0.04

0.02
0.8
0.00
0.6 -0.02
y(t)

0.4 0.30 0.32 0.34 0.36 0.38 0.40

Exact
0.2 M=1

M=2
0.0 M=3

0.0 0.2 0.4 0.6 0.8 1.0

t

Figure 1. (Example 2) Comparison between the exact solution and numerical solutions with M =
1, 2, 3.

Table 1. (Example 2) Absolute errors at some selected points with different M.

t M=2 M=6 M = 10
0.2 5.69 × 10−3 9.75 × 10−6 8.06 × 10−7
0.4 2.34 × 10−3 8.02 × 10−6 6.34 × 10−7
0.6 2.78 × 10−3 7.03 × 10−6 5.53 × 10−7
0.8 2.52 × 10−3 5.97 × 10−6 4.59 × 10−7
1.0 1.66 × 10−2 2.89 × 10−5 1.95 × 10−6

Example 3. Consider the following variable-order FDE taken from [42]:

C sin(t)
0 Dt y ( t ) + y ( t ) + e t y ( t5 ) = g ( t ), 0 < t ≤ 1,
(28)
y(0) = 0,
Fractal Fract. 2021, 5, 219 11 of 16

where
Γ (4) Γ (3)
g(t) = t3−sin(t) + t2−sin(t) + et (t15 + t10 ) + t3 + t2 .
Γ(4 − sin(t)) Γ(3 − sin(t))

The exact solution of this problem is y(t) = t3 + t2 . To solve Problem (28), we applied the
method with M = 1 and M = 2. The numerical solution obtained with M = 1, together with the
exact solution are plotted in Figure 2. With M = 2, according to the method described in Section 4,
we set n = max0<t≤1 {dsin(t)e} = 1. Therefore, by assuming that

y 0 ( t ) = A T B ( t ),

and using the initial condition, we have:

C sin(t) 1−sin(t)
y(t) = A T Pt1 B(t), 0 Dt y(t) = A T Pt B ( t ), y(t5 ) = A T Pt15 B(t5 ).

By substituting these approximations into (28) and using the collocation points t0 = 14 ,
t1 = 21 , and t2 = 34 , we obtain a system of three nonlinear algebraic equations in terms of the
elements of the vector A. By solving the resulting system, one obtains:

a0 = 2, a1 = 5, a2 = 3.

Finally, using these values, we obtain:

  
t 0 0 1
− 4t t
t − 12  = t3 + t2 ,
 
y(t) = 2 5 3 
2 0 
t
36 − 6t t
3
2
t −t+ 1
6

which is the exact solution.

2.0

1.5

Exact

M=1
y(t)

1.0

0.5

0.0

0.0 0.2 0.4 0.6 0.8 1.0

t

Figure 2. (Example 3) Comparison between the exact solution and numerical solution with M = 1.

Example 4. Let us now consider the fractional pantograph differential equation:

C α(t)
0 Dt y(t) + y(t) − 0.1y(0.2t) = −0.1e−0.2t , 0 < t ≤ 1, 0 < α(t) ≤ 1,
(29)
y(0) = 1.

The exact solution of this problem, when α(t) = 1, is y(t) = e−t . By choosing M = 1 and
α(t) = 1, we set:
y 0 ( t ) = A T B ( t ).
Then, by considering the initial condition, we have:

y(t) = A T Pt1 B(t) + 1, y(0.2) = A T P0.2t

1
B(0.2t) + 1.
Fractal Fract. 2021, 5, 219 12 of 16

By substituting these approximations into (29), and using the collocation points

1 2
t0 = , t1 = ,
3 3
we obtain
a0 = −0.620328, a1 = 0.621053,
which gives
  
t 0 1
+ 1 = 0.310526t2 − 0.930854t + 1.
 
y(t) = −0.620328 0.621053
− 4t t
2 t− 1
2

This approximate solution and the exact solution to the problem, corresponding to α = 1, are
displayed in Figure 3. By computing the L2 -norm of the error for this approximation, we have:

e−t − (0.310526t2 − 0.930854t + 1) = 6.29 × 10−3 ,

2

which shows that the method gives a high-accuracy approximate solution, even with a small number
of basis functions. A comparison of the absolute errors, obtained by the proposed method with
M = 6, 8, 10 at some selected points, with the results proposed in [43], using modified hat functions,
and those of [44], using Bernoulli wavelets, are reported in Table 2. From this table, it is seen that
our method gave more accurate results with a smaller number of basis functions when compared to
previous methods. Moreover, the approximate solutions obtained with M = 2 and different α(t),
along with the exact solution of corresponding first-order equation, are given in Figure 4. This
figure shows that the numerical solution is close to the exact solution for the case α(t) = 1 when
α(t) is close to one.

1.0

0.9
Exact
0.8 M=1

0.7
y(t)

0.6

0.5

0.4

0.0 0.2 0.4 0.6 0.8 1.0

t

Figure 3. (Example 4) Comparison between exact and numerical solutions with M = 1 and α(t) = 1.

Table 2. (Example 4) Comparison of absolute errors, at some selected grid points, with α(t) = 1.

Method of [43] Method of [44] Present Method

t n = 64 k = 2, M = 6 M=6 M=8 M = 10
2−2 1.18 × 10−9 1.05 × 10−8 8.61 × 10−9 1.37 × 10−11 5.56 × 10−13
2−3 5.39 × 10−10 5.79 × 10−9 1.01 × 10−8 1.57 × 10−11 4.25 × 10−13
2−4 1.17 × 10−9 2.00 × 10−8 9.30 × 10−9 1.59 × 10−11 2.42 × 10−13
2−5 5.34 × 10−10 3.70 × 10−9 6.47 × 10−9 1.21 × 10−11 1.29 × 10−13
2−6 2.27 × 10−9 2.03 × 10−8 3.83 × 10−9 7.58 × 10−12 6.72 × 10−14
Fractal Fract. 2021, 5, 219 13 of 16

1.0

Exact
0.9
α(t)=0.7+0.3sin2 (t)
0.8 α(t)=0.8+0.2t

α(t)=0.9+0.1sin(t)
0.7
y(t)
0.6

0.5

0.4

0.0 0.2 0.4 0.6 0.8 1.0

t

Figure 4. (Example 4) Approximate solutions with M = 2 and different α(t), together with the exact
solution for α(t) = 1.

Example 5. As our last example, consider the following variable-order FDE found in [45]:

Γ (1 − α ( t ), t )
 
C α(t) 0 t
0 Dt y(t) + 3y (t) − y(t) = e 3 − , 0 < t ≤ 1,
Γ(1 − α(t))

where α(t) = 0.25(1 + cos2 (t)) and the initial condition is y(0) = 1. The exact solution is
y(t) = et . We solved this problem using the proposed method in this paper with different values of
M. A comparison of the absolute errors at some selected points, obtained by our method and the one
of [45], based on a class of Lagrange polynomials, is given in Table 3.

Table 3. (Example 5) Comparison of the absolute errors at some selected grid points.

Method of [45] Present Method

t M=6 M = 10 M=6 M=8 M = 10
0.1 8.66 × 10−9 1.04 × 10−12 2.56 × 10−8 4.12 × 10−11 4.40 × 10−14
0.3 1.60 × 10−8 4.57 × 10−14 2.43 × 10−8 3.92 × 10−11 4.23 × 10−14
0.5 2.49 × 10−8 2.82 × 10−11 2.44 × 10−8 3.93 × 10−11 4.24 × 10−14
0.7 4.19 × 10−8 3.12 × 10−11 2.47 × 10−8 3.98 × 10−11 4.29 × 10−14
0.9 5.93 × 10−8 1.46 × 10−10 2.56 × 10−8 4.14 × 10−11 4.43 × 10−14

7. Concluding Remarks
Many researchers have employed fractional differential equations (FDEs) in order to
model and analyze various scientific phenomena. Typically, such FDEs do not have known
analytical solutions, and approximate and numerical approaches have to be applied [46].
Here, a new numerical method, based on Bernoulli polynomials, was presented for solv-
ing multiterm variable-order fractional differential equations. The operational matrix of
variable-order fractional integration for the Bernoulli basis functions was introduced, which
is a lower triangular matrix and helps to reduce the computational effort of the method.
Our scheme uses this matrix to give some approximations of the unknown solution of the
problem and its variable-order fractional derivatives in terms of the Bernoulli functions.
Substituting these approximations into the equation and using some collocation points al-
lowed us to reduce the problem to a system of nonlinear algebraic equations, which greatly
simplifies the problem. An error estimate of the method was proven, and the applicability
of our method was illustrated by solving five illustrative examples. The obtained results
confirmed the efficiency, accuracy, and high performance of our technique, when compared
with the state-of-the-art numerical schemes available in the literature. We emphasize that
the accuracy of our method is preserved even when the solution of the problem is not
7
infinitely differentiable. This can be observed in Example 2, whose exact solution is t 2 .
Fractal Fract. 2021, 5, 219 14 of 16

We used a variable-order definition where the operator has no order memory, a so-
called type-I operator. As future work, it would be interesting to extend the proposed
numerical method to approximate variable-order derivatives with weak (type-II) and strong
(type-III) variable-order definitions [18,47]. Other interesting lines of research include the
stability analysis of the proposed numerical method and its application to different areas
in science and engineering, for example in structural mechanics.

Author Contributions: Conceptualization, S.N., P.M.L., and D.F.M.T.; methodology, S.N., P.M.L.,
and D.F.M.T.; software, S.N.; validation, S.N., P.M.L., and D.F.M.T.; formal analysis, S.N., P.M.L., and
D.F.M.T.; investigation, S.N., P.M.L., and D.F.M.T.; writing—original draft preparation, S.N., P.M.L.,
and D.F.M.T.; writing—review and editing, S.N., P.M.L., and D.F.M.T.; visualization, S.N. All authors
have read and agreed to the published version of the manuscript.
Funding: This research was funded by Fundação para a Ciência e a Tecnologia (FCT, the Portuguese
Foundation for Science and Technology) through CEMAT, Grant Number UIDB/04621/2020 (P.M.L.),
and CIDMA, Grant Number UIDB/04106/2020 (D.F.M.T.).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Acknowledgments: The authors are grateful to the three anonymous reviewers for several construc-
tive remarks, questions, and suggestions.
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the
design of the study; in the collection, analyses, or interpretation of the data; in the writing of the
manuscript; nor in the decision to publish the results.

References
1. Almeida, R.; Tavares, D.; Torres, D.F.M. The Variable-Order Fractional Calculus of Variations; Springer: Cham, Switzerland, 2019.
https://doi.org/10.1007/978-3-319-94006-9.
2. Samko, S.G.; Kilbas, A.A.; Marichev, O.I. Fractional Integrals and Derivatives; Translated from the 1987 Russian Original; Gordon
and Breach: Yverdon, Switzerland, 1993.
3. Podlubny, I. Fractional Differential Equations; Academic Press: San Diego, CA, USA, 1999.
4. Kilbas, A.A.; Srivastava, H.M.; Trujillo, J.J. Theory and Applications of Fractional Differential Equations; Elsevier: Amsterdam,
The Netherlands, 2006.
5. Mainardi, F. Fractional Calculus and Waves in Linear Viscoelasticity; Imperial College Press: London, UK, 2010. https://doi.org/10.1
142/9781848163300.
6. Tuan, N.H.; Ganji, R.M.; Jafari, H. A numerical study of fractional rheological models and fractional Newell-Whitehead-Segel
equation with nonlocal and nonsingular kernel. Chin. J. Phys. 2020, 68, 308–320. https://doi.org/10.1016/j.cjph.2020.08.019.
7. Samko, S.G.; Ross, B. Integration and differentiation to a variable fractional order. Integr. Transf. Spec. Funct. 1993, 1, 277–300.
https://doi.org/10.1080/10652469308819027.
8. Odzijewicz, T.; Malinowska, A.B.; Torres, D.F.M. Noether’s theorem for fractional variational problems of variable order. Cent.
Eur. J. Phys. 2013, 11, 691–701. https://doi.org/10.2478/s11534-013-0208-2.
9. Chen, S.; Liu, F.; Burrage, K. Numerical simulation of a new two-dimensional variable-order fractional percolation equation in
non-homogeneous porous media. Comput. Math. Appl. 2014, 67, 1673–1681. https://doi.org/10.1016/j.camwa.2013.01.023.
10. Coimbra, C.F.M.; Soon, C.M.; Kobayashi, M.H. The variable viscoelasticity operator. Ann. Phys. 2005, 14, 378–389. https:
//doi.org/10.1002/andp.200410140.
11. Odzijewicz, T.; Malinowska, A.B.; Torres, D.F.M. Fractional Variational Calculus of Variable Order. In Advances in Harmonic
Analysis and Operator Theory. Operator Theory: Advances and Applications; Almeida, A., Castro, L., Speck, F.O., Eds.; Birkhäuser:
Basel, Switzerland; 2013; Volume 229, 291–301. https://doi.org/10.1007/978-3-0348-0516-2_16.
12. Ostalczyk, P.W.; Duch, P.; Brzeziński, D.W.; Sankowski, D. Order functions selection in the variable fractional-order PID
controller. Advances in Modelling and Control of Non-integer-Order Systems. Lect. Notes Electr. Eng. 2015, 320, 159–170.
https://doi.org/10.1007/978-3-319-09900-2_15.
13. Rapaić, M.R.; Pisano, A. Variable-order fractional operators for adaptive order and parameter estimation. IEEE Trans. Autom.
Contr. 2014, 59, 798–803. https://doi.org/10.1109/TAC.2013.2278136.
14. Patnaik, S.; Semperlotti, F. Variable-order particle dynamics: formulation and application to the simulation of edge dislocations.
Phil. Trans. R. Soc. A 2020, 378, 20190290. https://doi.org/10.1098/rsta.2019.0290.
Fractal Fract. 2021, 5, 219 15 of 16

15. Blaszczyk, T.; Bekus, K.; Szajek, K.; Sumelka, W. Approximation and application of the Riesz-Caputo fractional derivative of
variable order with fixed memory. Meccanica 2021, in press. https://doi.org/10.1007/s11012-021-01364-w.
16. Di Paola, M.; Alotta, G.; Burlon, A.; Failla, G. A novel approach to nonlinear variable-order fractional viscoelasticity. Phil. Trans.
R. Soc. A 2020, 378, 20190296. https://doi.org/10.1098/rsta.2019.0296.
17. Burlon, A.; Alotta, G.; Di Paola, M.; Failla, G. An original perspective on variable-order fractional operators for viscoelastic
materials. Meccanica 2021, 56, 769–784. https://doi.org/10.1007/s11012-021-01316-4.
18. Patnaik, S.; Jokar, M.; Semperlotti, F. Variable-order approach to nonlocal elasticity: theoretical formulation, order identification
via deep learning, and applications. Comput. Mech. 2021, in press. https://doi.org/10.1007/s00466-021-02093-3.
19. El-Sayed, A.A.; Agarwal, P. Numerical solution of multiterm variable-order fractional differential equations via shifted Legendre
polynomials. Math. Meth. Appl. Sci. 2019, 42, 3978–3991. https://doi.org/10.1002/mma.5627.
20. Wang, L.F.; Ma, Y.P.; Yang, Y.Q. Legendre polynomials method for solving a class of variable order fractional differential equation.
CMES-Comp. Model. Eng. 2014, 101, 97–111. https://doi.org/10.3970/cmes.2014.101.097.
21. Chen, Y.M.; Wei, Y.Q.; Liu, D.Y.; Yu, H. Numerical solution for a class of nonlinear variable order fractional differential equations
with Legendre wavelets. Appl. Math. Lett. 2015, 46, 83–88. https://doi.org/10.1016/j.aml.2015.02.010.
22. Liu, J.; Li, X.; Wu, L. An operational matrix of fractional differentiation of the second kind of Chebyshev polynomial for solving
multiterm variable order fractional differential equation. Math. Probl. Eng. 2016, 2016, 7126080. https://doi.org/10.1155/2016/7
126080.
23. Nagy, A.M.; Sweilam, N.H.; El-Sayed, A.A. New operational matrix for solving multiterm variable order fractional differential
equations. J. Comput. Nonlinear Dyn. 2018, 13, 11001–11007. https://doi.org/10.1115/1.4037922.
24. Chen, Y.M.; Liu, L.Q.; Li, B.F.; Sun, Y. Numerical solution for the variable order linear cable equation with Bernstein polynomials.
Appl. Math. Comput. 2014, 238, 329–341. https://doi.org/10.1016/j.amc.2014.03.066.
25. Tavares, D.; Almeida, R.; Torres, D.F.M. Caputo derivatives of fractional variable order: Numerical approximations. Commun.
Nonlinear Sci. 2016, 35, 69–87. https://doi.org/10.1016/j.cnsns.2015.10.027.
26. Shen, S.; Liu, F.; Chen, J.; Turner, I.; Anh, V. Numerical techniques for the variable order time fractional diffusion equation. Appl.
Math. Comput. 2012, 218, 10861–10870. https://doi.org/10.1016/j.amc.2012.04.047.
27. Hammouch, Z.; Yavuz, M.; Özdemir, N. Numerical solutions and synchronization of a variable-order fractional chaotic system.
Math. Model. Numer. Simul. Appl. 2021, 1, 11–23. https://doi.org/10.53391/mmnsa.2021.01.002.
28. Ganji, R.M.; Jafari, H.; Baleanu, D. A new approach for solving multi variable orders differential equations with Mittag-Leffler
kernel. Chaos Solitons Fractals 2020, 130, 109405. https://doi.org/10.1016/j.chaos.2019.109405.
29. Eghbali, A.; Johansson, H.; Saramäki, T. A method for the design of Farrow-structure based variable fractional-delay FIR filters.
Signal Process. 2013, 93, 1341–1348. https://doi.org/10.1016/j.sigpro.2012.11.010.
30. Yu, C.; Teo, K.L.; Dam, H.H. Design of all pass variable fractional delay filter with signed powers-of-two coefficients. Signal
Process. 2014, 95, 32–42. https://doi.org/10.1016/j.sigpro.2013.08.005.
31. Cooper, G.R.J.; Cowan, D.R. Filtering using variable order vertical derivatives. Comput. Geosci. 2004, 30, 455–459. https:
//doi.org/10.1016/j.cageo.2004.03.001.
32. Tseng, C.-C. Design of variable and adaptive fractional order FIR differentiators. Signal Process. 2006, 86, 2554–2566. https:
//doi.org/10.1016/j.sigpro.2006.02.004.
33. Bhrawy, A.H.; Tohidi, E.; Soleymani, F. A new Bernoulli matrix method for solving high-order linear and nonlinear Fredholm
integro-differential equations with piecewise intervals. Appl. Math. Comput. 2012, 219, 482–497. https://doi.org/10.1016/j.amc.
2012.06.020.
34. Tohidi, E.; Bhrawy, A.H.; Erfani, K. A collocation method based on Bernoulli operational matrix for numerical solution of
generalized pantograph equation. Appl. Math. Model. 2013, 37, 4283–4294. https://doi.org/10.1016/j.apm.2012.09.032.
35. Toutounian, F.; Tohidi, E. A new Bernoulli matrix method for solving second order linear partial differential equations with the
convergence analysis. Appl. Math. Comput. 2013, 223, 298–310. https://doi.org/10.1016/j.amc.2013.07.094.
36. Bazm, S. Bernoulli polynomials for the numerical solution of some classes of linear and nonlinear integral equations. J. Comput.
Appl. Math. 2015, 275, 44–60. https://doi.org/10.1016/j.cam.2014.07.018.
37. Keshavarz, E.; Ordokhani, Y.; Razzaghi, M. A numerical solution for fractional optimal control problems via Bernoulli polynomials.
J. Vib. Control 2016, 22, 3889–3903. https://doi.org/10.1177/1077546314567181.
38. Costabile, F.; Dellaccio, F.; Gualtieri, M.I. A new approach to Bernoulli polynomials. Rend. Mat. Ser. VII 2006, 26, 1–12.
39. Arfken, G. Mathematical Methods for Physicists, 3rd ed.; Academic Press: San Diego, CA, USA, 1985.
40. Mason, J.C.; Handscomb, D.C. Chebyshev Polynomials; CRC Press LLC: Boca Raton, FL, USA, 2003.
41. Hassani, H.; Dahaghin, M.S.; Heydari, H. A new optimized method for solving variable-order fractional differential equations. J.
Math. Ext. 2017, 11, 85–98.
42. Li, X.; Li, H.; Wu, B. A new numerical method for variable order fractional functional differential equations. Appl. Math. Lett.
2017, 68, 80–86. https://doi.org/10.1016/j.aml.2017.01.001.
43. Nemati, S.; Lima, P.; Sedaghat, S. An effective numerical method for solving fractional pantograph differential equations using
modification of hat functions. Appl. Numer. Math. 2018, 131, 174–189. https://doi.org/10.1016/j.apnum.2018.05.005.
44. Rahimkhani, P.; Ordokhani, Y.; Babolian, E. A new operational matrix based on Bernoulli wavelets for solving fractional delay
differential equations. Numer. Algorithms 2017, 74, 223–245. https://doi.org/10.1007/s11075-016-0146-3.
Fractal Fract. 2021, 5, 219 16 of 16

45. Sabermahani, S.; Ordokhani, Y.; Lima, P.M. A novel Lagrange operational matrix and Tau-collocation method for solving
variable-order fractional differential equations. Iran J. Sci. Technol. Trans. Sci. 2020, 44, 127–135. https://doi.org/10.1007/s40995
-019-00797-z.
46. Nikan, O.; Jafari, H.; Golbabai, A. Numerical analysis of the fractional evolution model for heat flow in materials with memory.
Alex. Eng. J. 2020, 59, 2627–2637. https://doi.org/10.1016/j.aej.2020.04.026.
47. Lorenzo, C.F.; Hartley, T.T. Variable order and distributed order fractional operators. Nonlinear Dyn. 2002, 29, 57–98. https:
//doi.org/10.1023/A:1016586905654.