2018.proceedings Cmmse 7

Proceedings of the 18th International Conference
on Computational and Mathematical Methods

in Science and Engineering, CMMSE 2018
July 9–14, 2018.
Supratransmission in nonlinear fractional wave systems
Luis E. Piña-Villalpando1 , J. E. Macı́as-Dı́az2 and Evguenii Kourmychev1

1 Departamento de Ciencias Exactas y Tecnologı́a, Centro Universitario de los Lagos,
Universidad de Guadalajara, Mexico
2 Departamento de Matemáticas y Fı́sica, Universidad Autónoma de Aguascalientes
emails: lepina@correo.uaa.mx, jemacias@correo.uaa.mx, hvargas@culagos.udg.mx
Abstract
In this work, we will investigate numerically the presence of nonlinear supratrans-

mission in various fractional systems. We will consider the continuous forms of the clas-
sical Fermi–Pasta–Ulam chains, arrays of Josephson junctions and double sine-Gordon
chains. We will employ a numerical technique with consistency properties in the energy
domain, and our simulations will be carried out using an efficient computer implemen-
tation of that technique. As a result, we will show numerically that supratransmission
is present in such continuous fractional systems.
Key words: nonlinear supratransmission, nonlinear media, fractional systems, nu-
merical simulations
MSC 2000: 35B40, 35B65, 35B09
1 Definitions
Let p ∈ N, T ∈ R+ and γ ∈ R+ ∪ {0}. Suppose that ai < bi and αi ∈ (0, 1) ∪ (1, 2], for each
1 ≤ i ≤ p. Let
p
Y
B = (ai , bi ) ⊆ Rp , (1)
i=1
Ω = B × (0, T ) ⊆ Rp+1 . (2)
We use Lx,2 (Ω) to denote the set of f : Ω → R satisfying f (·, t) ∈ L2 (B) for each t ∈ [0, T ].
If f, g ∈ Lx,2 (Ω) then their inner product is the function of t given by
Z
hf, gix = f (x, t)g(x, t)dx, ∀t ∈ [0, T ]. (3)
B
CMMSE
c ISBN: 978-84-697-7861-6
SUPRATRANSMISSION IN FRACTIONAL SYSTEMS
p
The Euclidean norm of f ∈ Lx,2 (Ω) is the function of t defined by kf kx,2 = hf, f i. The
set of all f : Ω → R such that f (·, t) ∈ L1 (B) for each t ∈ [0, T ] will be denoted by Lx,1 (Ω),
and their respective norms are defined by
Z
kf kx,1 = |f (x, t)|dx, ∀t ∈ [0, T ]. (4)
B
Let u : Ω → R be a function and convey that x = (x1 , x2 , . . . , xp ) ∈ Rp . If 1 ≤ i ≤ p

then the Riesz derivatives of order αi of u with respect to the variable xi evaluated at
(x, t) ∈ Ω, is given by
Z bi
∂ αi u −1 ∂n u(x1 , . . . , xi−1 , ξ, xi+1 , . . . , xp , t)
(x, t) = dξ, (5)
∂|xi |αi 2 cos( πα
2
i
)Γ(n − α i ) ∂xn
i ai |xi − ξ|αi +1−n
where n − 1 < αi ≤ n. For convenience, denote the Riesz derivative of order αi with respect
to xi by Dxαii . The unique square root of −Dxαii will be denoted here by Ξαxii , and it satisfies
the following for any functions u and v[1, 2]:
∂ αi u

α αi

− , v = Ξx
i
u, Ξx v . (6)
∂|xi |αi x
i i x
2 Fermi–Pasta–Ulam model
Let Ω be equal to either [a, b] or [a, ∞) and suppose that and γ are nonnegative numbers.
Assume that p ≥ 1 and let u = u(x, t) be a real function of (x, t) ∈ Ω × [0, ∞). A continuous
form of the β-Fermi–Pasta–Ulam chain is provided by the nonlinear partial differential
equation
p
∂2u ∂2u

∂u ∂u
(x, t) = (x, t) 1 + (x, t) − γ (x, t), ∀(x, t) ∈ Ω × [0, ∞). (7)
∂t2 ∂x2 ∂x ∂t
Clearly, the Hamiltonian and the total energy of the undamped system are, respectively,
" 2 2 # p+2
1 ∂u ∂u ∂u
H(x, t) = (x, t) + (x, t) + (x, t) , (8)
2 ∂t ∂x (p + 2)(p + 1) ∂x
Z b
E(t) = H(x, t)dx, (9)
a
for each (x, t) ∈ Ω × [0, ∞).
In order to provide a fractional extension of (7), we define u(x, t) = 0 for each x ∈
/Ω
and consider the Riesz fractional derivative of order α ∈ (1, 2), which is given by the one-
dimensional operator
Z ∞
∂αu −1 ∂2 u(ξ, t)
= dξ, ∀(x, t) ∈ Ω × [0, ∞). (10)
∂|x|α 2 cos( πα
2 )Γ(2 − α) ∂x2
−∞ |x − ξ|α−1
CMMSE
c ISBN: 978-84-697-7861-6
L. E. PIÑA ET AL.
Here Γ(z) represents the gamma function, for each z ∈ R \ {n : −n ∈ N or n = 0}. Using
this notation, a fractional form of (7) is readily found when replacing the second-order
spatial derivative by the Riesz space-fractional operator. In the presence of a physically
meaningful formulation of the Euler–Lagrange formality for fractional systems (see [?]), we
propose the following Hamiltonian for the undamped scenario:
" 2 # p+2
∂αu

1 ∂u ∂u
H(x, t) = (x, t) + u(x, t) (x, t) + (x, t) , (11)
2 ∂t ∂|x|α (p + 2)(p + 1) ∂x
for each (x, t) ∈ Ω × [0, ∞). Under these circumstances, a functional for the total energy of
the system would be provided by (9).
3 Double sine-Gordon model

Suppose that G : R → R is a nonnegative function, and that φ, ψ : B → R satisfy φ(x) =
ψ(x) = 0 for each x ∈ ∂B. Throughout, we let the function u : Ω → R satisfy the initial-
boundary-value problem
p
∂2u X ∂ αi u ∂u
2
(x, t) − α
(x, t) + γ (x, t) + G0 (u(x, t)) = 0, ∀(x, t) ∈ Ω,
∂t ∂|xi | i ∂t
i=1
∀x ∈ B,

 u(x, 0) = φ(x), (12)
∂u

such that (x, 0) = ψ(x), ∀x ∈ B,
 ∂t

u(x, t) = 0, ∀(x, t) ∈ ∂B × (0, T ).
Classical diffusion is obtained if all αi = 2, in which case the derivatives are the usual
differential operators. Otherwise, we will say that the diffusion is anomalous.
The literature has proposed various functionals to calculate the energy of undamped
systems governed by (12) in one spatial dimension[3, 4]. We will use the following function,
which is a dimensional extension of that used in[5]:
2 p
1 ∂u 1 X Ξαx i u 2 + kG(u)k , ∀t ∈ (0, T ).

E(t) = + i x,2 x,1 (13)
2 ∂t x,2 2

i=1
Obviously, the corresponding energy density is defined by
2 p
∂ αi u

1 ∂u 1X
H(x, t) = (x, t) − u(x, t) (x, t) + G(u(x, t)), (14)
2 ∂t 2 ∂|xi |αi
i=1
for each (x, t) ∈ Ω. Note that the function E is nonnegative. Also, if u is a solution of (12)
then E 0 (t) = −γkut k2x,2 , for each t ∈ (0, T ). Moreover,
Z t 2
∂u
E(t) − E(0) = −γ dt, ∀t ∈ (0, T ).
∂t (15)
0 x,2
CMMSE
c ISBN: 978-84-697-7861-6
SUPRATRANSMISSION IN FRACTIONAL SYSTEMS
Obviously, if γ = 0 then (12) is conservative.
4 Josephson transmission model

In this work, we will also consider a fractional generalization using discrete fractional oper-
ators. More concretely, we will investigate the model
ün (t) − c2 Dα un (t) + γn u̇n (t) + sin un (t) = µ, ∀n ∈ JN ,
∀n ∈ JN ,

u̇ (0) = un (0) = 0,
 n

φ(t) (16)
such that u0 (t) − u1 (t) = 2 , ∀t > 0,

 c
uN +1 (t) − uN (t) = 0, ∀t > 0,
It is important to mention that the undamped Riesz-fractional system (16) has the
following fractional Hamiltonian-like functional associated to the nth node, for each n ∈ JN :
1 2
u̇n (t) − c2 un (t)Dα un (t) + 1 − cos un (t)

Hn (t) = (17)
2
(see [3]). As a consequence, the energy functional associated to the undamped form of
(16) is given by (13). Moreover, a useful quantity in our investigation is the total energy
administered into the system over a period of time T > 0, which is provided by
Z T
ET = E(t)dt. (18)
0
References
[1] Avner Friedman. Foundations of modern analysis. Courier Corporation, New York,
1970.
[2] Nick Laskin. Fractional Schrödinger equation. Physical Review E, 66(5):056108, 2002.
[3] Georgy Alfimov, Teresa Pierantozzi, and Luis Vázquez. Numerical study of a fractional
sine-Gordon equation. In A. Le Mehuate, J. A. Tenreiro Machado, L. C. Trigeassou,
and J. Sabatier, editors, Fractional differentiation and its applications, volume 4, pages
153–162, Bordeaux, France, July 2004. Proceedings of the IFAC-FDA Workshop.
[4] Vasily E Tarasov and George M Zaslavsky. Conservation laws and Hamiltons equations
for systems with long-range interaction and memory. Communications in Nonlinear
Science and Numerical Simulation, 13(9):1860–1878, 2008.
[5] J. E. Macı́as-Dı́az, A. S. Hendy, and R. H. de Staelen. A pseudo energy-invariant method
for relativistic wave equations with Riesz space-fractional derivatives. Computer Physics
Communications, (submitted for publication), 2017.
CMMSE
c ISBN: 978-84-697-7861-6

2018.proceedings Cmmse 7

Uploaded by

Copyright:

Available Formats

2018.proceedings Cmmse 7

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2018.proceedings Cmmse 7

Uploaded by

Copyright:

Available Formats

Proceedings of the 18th International Conference

on Computational and Mathematical Methods

Supratransmission in nonlinear fractional wave systems

Luis E. Piña-Villalpando1 , J. E. Macı́as-Dı́az2 and Evguenii Kourmychev1

In this work, we will investigate numerically the presence of nonlinear supratrans-

Let u : Ω → R be a function and convey that x = (x1 , x2 , . . . , xp ) ∈ Rp . If 1 ≤ i ≤ p

3 Double sine-Gordon model

Obviously, if γ = 0 then (12) is conservative.

4 Josephson transmission model

You might also like