Optik - International Journal for Light and Electron Optics 206 (2020) 164356

Contents lists available at ScienceDirect

Optik
journal homepage: www.elsevier.com/locate/ijleo

Original research article

Invariant analysis and conservation laws of time fractional

T
Schrödinger equations
Q. Hussaina, F.D. Zamana, A.H. Karab,*
a
Abdus Salaam School of Mathematical Sciences, Government College University, Lahore, Pakistan
b
School of Mathematics, University of the Witwatersrand, Johannesburg, Wits 2001, South Africa

A R T IC LE I N F O ABS TRA CT

Keywords: We study the invariance properties (symmetries) and conservation laws of the fractional time
Symmetries version of the nonlinear Schrödinger equation with power law nonlinearity iytα + yxx + y|y|n = 0,
Conservation laws for 0 < α < 1, using some recently developed approaches. We will show that the all important
Time fractional energy conservation due time invariance is lost due to some built in approach that the theory
Schrodinger equations
necessitates

1. Introduction

The nonlinear Schrödinger equation (NLSE) with cubic nonlinearity is a well studied partial differential equation (pde) with wide
applications including optics and various areas of mathematical physics. The indepth studies follow their impact on certain wave
phenomena and the methods employed are varied; soliton, travelling wave solutions to conservation laws (to name a few, [1–5]).
However, the developing area of time fractional approaches to the NLSE has been considered but specific to some methods [6–8]).
For methods involving invariances and symmetries, there is seems be to no such work. To this end, in this paper, we consider a
nonlinear time-fractional Schrödinger equation (TFS) with cubic nonlinearity given by

iytα + yxx + y|y|2 = 0, 0 < α < 1, (1)

where ytα = Dtα y is a fractional derivative of function y with respect to t of order α, 0 < t < T (T ≤∞), x ∈ Ω ⊂ R and q is complex
valued. Here, we will take ytα to be the Riemann-Liouville left-sided time-fractional 0Dtα given by
α
0 Dt y = Dtn ( 0Itn − α y ), (2)

where Dt is the operator of total differentiation with respect to t, n = α + 1 and 0Itn − α y is the left-sided time-fractional integral of
order n − α defined by

1 t y (θ , x )
( 0Itn − α y )(t , x ) =
Γ(n − α )
∫0 (t − θ)1 − n + α
dθ ,
(3)

where Γ(.) is the Gamma function.

We note that the Caputo left-sided time- fractional derivative C0 Dtα y is given by
C α
0 Dt y = 0Itn − α (Dtn y ). (4)

⁎
Corresponding author.
E-mail address: abdul.kara@wits.ac.za (A.H. Kara).

https://doi.org/10.1016/j.ijleo.2020.164356
Received 3 February 2020; Accepted 3 February 2020
0030-4026/ © 2020 Elsevier GmbH. All rights reserved.
Q. Hussain, et al. Optik - International Journal for Light and Electron Optics 206 (2020) 164356

For details on definitions and results based on fractional equations, we mention some references, viz., [9–12]. However, we recall
some pertinent differential operators which details may be found in [13] and references therein. These are the Erdélyi-Kober frac-
tional differential operator
m−1
1 d
( ζδ, α)(z ) = ∏ (ζ + j − z )( ζδ+ α, m − α h)(z ), z > 0, δ > 0, m = [α ] + 1,
j=0
δ dz (5)

where
1 ∞ 1
( ζδ, α h)(z ) =
Γ(α )
∫1 (p − 1)α − 1p−(ζ + α ) h (zp δ )dp .
(6)

2. Invariance and Lie symmetries

If y = u + iv, (1) becomes

utα + vxx + v (u2 + v 2) = 0,

− vtα + u xx + u (u2 + v 2) = 0. (7)

In what follows, we will analyse the invariance (Lie point symmetries) and conservation laws of the TFS with cubic nonlinearity via
the system (7).
The vector field
∂ ∂ ∂ ∂
X = τ (t , x , u, v ) + ξ (t , x , u, v ) + η (t , x , u, v ) + ϕ (t , x , u, v ) ,
∂t ∂x ∂u ∂v (8)

is a Lie point symmetry generator of the system (7) if

X [α,2] [utα + vxx + v (u2 + v 2)] = 0,

X [α,2] [−vtα + u xx + u (u2 + v 2)] = 0 (9)

along the solutions of (7) with in (9) prolonged accordingly as

∂ ∂ ∂ ∂ ∂ ∂ ∂ ∂
X [α,2] = τ +ξ +η +ϕ + ηα, t α + η x + ϕ α, t α + ϕ x
∂t ∂x ∂u ∂v ∂ut ∂u x ∂vt ∂vx (10)

∂ ∂
+ η xx + ϕ xx ,
∂u xx ∂vxx (11)

where

ηα, t = Dtα η + ξDtα (u x ) − Dtα (ξu x ) + τDtα (ut ) − Dtα (τt ),

ηx= Dx η − u x Dx ξ − u t Dx τ ,
ϕ α, t = Dtα ϕ + ξDtα (vx ) − Dtα (ξvx ) + τDtα (vt ) − Dtα (τt ),
ϕ x= Dx ϕ − vx Dx ξ − vt Dx τ ,
η xx = Dx η x − u xx Dx ξ − u xt Dx τ ,
ϕ xx= Dx ϕ x − vxx Dx ξ − vxt Dx τ , (12)

where
∞ ∞
∂αη ∂αu ∂αη ∂αv ∂αη ∂nηu α − n ∂nηv α − n
Dtα η= α
+ (ηu α − u αu ) + (ηv α − v αv ) + ∑ (αn) Dt u + ∑ (αn) Dt v
∂t ∂t ∂t ∂t ∂t n=1
∂t n n=1
∂t n
∞ ∞ ∞ ∞
1 t n−α ∂m ∂n − m + k
+ ∑∑∑∑ (αn)(nm)(kr )
k! Γ(n − α + 1)
(−u)r m (uk − r ) n − m + k η
∂t ∂t
n=1 n=1 n=1 n=1
∞ ∞ ∞ ∞
1 t n−α ∂m ∂n − m + k
+ ∑ ∑ ∑ ∑ (αn)(nm)(kr ) k! Γ(n − α + 1) (−v )r ∂t m (v k −r ) ∂t n−m +k η
n=1 n=1 n=1 n=1 (13)

and similarly for The operator Dx is the standard ‘total derivative operator’ with respect to x.
Dtα ϕ .
Now, (11) becomes (with some arbitrary k)

ηα, t + ϕ xx + kϕ (u2 + 3v 2) + 2kηuv = 0,

− ϕ α, t + η xx + kϕ (3u2 + v 2) + 2kϕ uv = 0 (14)

subject to (7). Expanding and separating by monomials (‘powers’ of the derivatives), we get two sets of systems of which the first is

2
Q. Hussain, et al. Optik - International Journal for Light and Electron Optics 206 (2020) 164356

ux : ϕxu = 0, vx : 2ϕxv − ξ xx = 0,
u x vx : ϕuv = 0, vx2: ϕvv = 0,
vxx : ϕv − 2ξ x − ηu + ατt = 0, u x2:ϕuu = 0,
u xx : ϕu + ηv = 0, vt u xx : τ = 0,
vx u xx : ξu = 0, vx vxx : ξ v = 0,
vt vxx : τv = 0, vtx : τx = 0 (15)

and

∂tα η − u∂tα ηu − v ∂tα ηv − kv(ηu − ατt )(u2 + v 2) + ηv ku(u2 + v 2)

+ ϕxx + kϕ (u2 + 3v 2) + 2kuvη = 0. (16)

Also, for n = 1, 2, 3…,

∞
Dtα − n u x : ∑ (αn) Dtn ξ = 0,
n=1
∞
∂nηu
Dtα − n u: ∑ [(αn) − (αn + 1) Dtn + 1 τ ] = 0,
n=1
∂t n
∞
∂nηv
Dtα − n v: ∑ (αn) = 0,
n=1
∂t n (17)

The second set is

ux : 2ηxu − ξ xx = 0, vx : ηxv = 0,
u x vx : ηuv = 0, vx2: ηvv = 0,
vxx : ηv + ϕu = 0, u x2: ηuu = 0,
u xx : ηu − 2ξ x − ϕv + ατt = 0, (18)

and

− ∂tα ϕ + v ∂tα ϕv + u∂tα ϕu − kv(ϕv − ατt )(u2 + v 2) + ϕu kv(u2 + v 2)

+ ηxx + kη (3u2 + v 2) + 2kuvϕ = 0, (19)

and with n = 1, 2, 3…,

∞
Dtα − n vx : ∑ (αn) Dtn ξ = 0,
n=1
∞
∂nϕv
Dtα − n v: ∑ [(αn) − (αn + 1) Dtn + 1 τ ] = 0,
n=1
∂t n
∞
∂nϕu
Dtα − n u: ∑ (αn) = 0.
n=1
∂t n (20)

In (17) and (20), the cases n = 1 and n = 2 are sufficient. After some detailed calculations and the requirement that τ|τ=0 = 0, we get
the generators
2
∂x , v ∂u − u∂v , x ∂x + t ∂t − u∂u − v ∂v .
α (21)

We note that the system is not invariant under time translation (energy not conserved) and, additionally, there is another loss of
symmetry from the well know case α = 1.

2.1. A discussion on reductions

Since the system (7) is not invariant in time t, the Riemann-Liouville fractional equation does not display travelling wave or steady
state solutions.
2 2
We present a summary of a reduction based on the scaling symmetry X = x ∂x + α t ∂t − u∂u − v ∂v (W u = −u − xux − α tut whose
invariants are obtained by the system of first-order odes
dx dt du dv
= 2 = = ,
x t −u −v (22)
α

viz.,

3
Q. Hussain, et al. Optik - International Journal for Light and Electron Optics 206 (2020) 164356

α 2 2
z = xt− 2 , f (z ) = t − α u (x , t ), g (z ) = t − α v (x , t ),

say. The analysis, further, requires the use of the operators defined in (5). After some detailed and tedious calculations, it can be
shown that the system (7) reduces to
1− α ,α
( 2/ α2 f )(z )= − g ″ (z ) − g (z )(g 2 (z ) + f 2 (z )),
1− α ,α
( 2/ α2 g )(z )= f ″ (z ) + f (z )(g 2 (z ) + f 2 (z )). (23)

3. Conservation laws

A vector (Tt, Tx) is a conserved vector (flow) corresponding to a conservation law of (7), if
Dt T t + Dx T x = 0 (24)
t x
along the solutions of the system (7). If DtT + DxT vanishes identically, i.e., irrespective of (7), we have a trivial conservation law.
The method we will employ to construct the conserved flows will be an adaptation of Ibragimov's formal Lagrangian approach ([14],
see also [15]) which depend on solutions of the adjoint equation/system (in this case of (7)).
The formal Lagrangian here is given by
Ł = p (t , x )(utα + vxx + v (u2 + v 2)) + q (t , x )(−vtα + u xx + u (u2 + v 2)) (25)
for which the action integral is
T
∫0 ∫Ω Ł(t, x, u, v, p, q, utα , ux , u xx , vtα, vx , vxx )dx dt (26)
and the Euler operator is given by
δŁ ∂Ł ∂Ł ∂Ł ∂Ł
= + (Dtα )* α − Dx + Dx Dx ,
δu ∂u ∂ut ∂u x ∂u xx
δŁ ∂Ł ∂Ł ∂Ł ∂Ł
= + (Dtα )* α − Dx + Dx Dx ,
δv ∂v ∂vt ∂vx ∂vxx (27)
where (Dtα )* is the adjoint, in the Frechet sense, of Dtα , so that
δŁ
= 2puv + q (3u2 + v 2) + (Dtα )*p + qxx ,
δu
δŁ
= 2quv + p (u2 + 3v 2) + (Dtα )*(−q) + pxx
δv (28)
δŁ δŁ
with δp
= utα
+ vxx + v (u2 + v 2) and = δq
−vtα
+ u xx + u (u2 + v 2) . The functions p(t, x) and q(t, x) are solutions of the adjoint system
of (7).
In order to construct the conserved flows, we require the No ether operator, which is discussed in detail in [16] in its role in
constructing conserved vectors. In its fractional setup, the operation of the Noether operator on the formal Lagrangian Ł lead to the
conserved flows [15]. Given a Lie point symmetry, vector field, of a time fractional system of pdes (u = u(t, x) and v = v (t , x ) )
∂ ∂ ∂ ∂
X=τ +ξ +η +ϕ ,
∂t ∂x ∂u ∂v (29)
the respective Noether operators for the Riemann-Liouville case are given by [15]
n−1
∂ ∂
 t= τ  + ∑ (−1)k 0Dtα−1−k (W u) Dtk ∂ ( α
− (−1)nJ (W u, Dtn α
)
k=0
D
0 t u ) ∂ ( 0 t u)
D
n−1
∂ ∂
+ ∑ (−1)k 0Dtα−1−k (W v ) Dtk ∂ ( α
− (−1)nJ (W v, Dtn
∂ ( 0Dtα v )
),
k=0 0 Dt v )

∂ ∂ ∂ ∂ ∂ ∂
 x= ξ  + W u ( − Dx ) + Dx (W u ) + Wv( − Dx ) + Dx (W v ) ,
∂u x ∂u xx ∂u xx ∂vx ∂vxx ∂vxx (30)
where  is the identity operator, W = η − utτ − uxξ and u
Wv = ϕ − vt τ − vx ξ are the characteristics of X,
1 t T f (θ , x ) g (μ, x )
J (f , g ) =
Γ(n − α )
∫0 ∫t (μ − θ)α + 1 − n
dμdθ
(31)
and n = [α] + 1 (since 0 < α < 1, here, n = 1). Then, the components of the conserved vectors are
T t =  tŁ, T x =  xŁ. (32)
∂Ł ∂Ł ∂Ł ∂Ł x
Since ∂ux
= ∂v x
= 0, ∂uxx
= q and ∂v xx
= q , T is straightforward, we get

4
Q. Hussain, et al. Optik - International Journal for Light and Electron Optics 206 (2020) 164356

T t = τ Ł + Dtα − 1 (W u ) p + J (W u, pt ) + Dtα − 1 (W v )(−q) + J (W v, −qt ). (33)

Thus, we have the following cases of non-trivial conserved flows (Tt, Tx) associated with symmetry X.
(i) X = −v ∂u + u∂v (W u = −v and W v = u )

T t = − vtα − 1 p + J (−v, pt ) + utα − 1 (−q) + J (u, −qt ),

T x = vqx − qvx − upx + pu x (34)
so that the total divergence is

Dt T t + Dx T x = pu xx − pDt (vtα − 1) − qvxx − qDt (utα − 1) − utα − 1 qt − upxx

− vtα − 1 pt + vq xx + Dt J (−v, pt ) + Dt J (u, −qt ). (35)
(ii) X = ∂x (W =− ux and
u
Wv = −vx ) - linear momentum

T t= − u xtα − 1 p + J (−u x , pt ) + vxtα − 1 q + J (−vx , −qt ),

T x= p [utα + v (u2 + v 2)] + q [−vtα + u (u2 + v 2)] − u x qx + vx px . (36)
∞ ∞
The momentum conserved quantity is given by ∫ viz., ∫
−∞
T t dx , [−u xtα − 1 p
+ J (−u x , pt ) +
−∞
vxtα − 1 q + J (−vx , −qt )]dx if |u| and |v| and
the derivatives of u and v converge.
2 2 2
(iii) X = x ∂x + α t ∂t − u∂u − v ∂v (W u = −u − xux − α tut and W v = −v − xvx − α tvt )
2
T t= t [p ((utα + vxx + v (u2 + v 2)) + q (−vtα + u xx + u (u2 + v 2))]
α
2 2
+ J (−u − xux − tut , pt ) + J (−v − xvx − tvt , −qt )
α α
2 2
+ p [−utα − 1 − xuαxt− 1 − Dtα − 1 (tut )] − q [−vtα − 1 − xv αxt− 1 − Dtα − 1 (tvt)],
α α
T x = x [p ((utα + vxx + v (u2 + v 2)) + q (−vtα + u xx + u (u2 + v 2))]
2 2
+ q (−2u x − tuxt) − px ( −v − xvx − tvt)
α α
2 2
+ p (−2vx − tvxt) − qx ( −u − xux − tut )
α α (37)

4. Discussion

We have shown that the well known Schrödinger equation has interesting invariance properties for the fractional time version. Its
conservation laws can be constructed using some recently developed approaches. Importantly, the energy conservation due time
invariance is lost and this is due to some built in approach that the theory necessitates. One can pursue a practical explanation for this
and check whether this is in line with alternative methods. Nevertheless, we have shown that by construing the standard Schrödinger
equations as systems, the methods of symmetries and conservation laws can be applied to systems in general. In particular, for the
general nonlinear Schrödinger equations iytα + yxx + y|y|n = 0 , or iytα + yxx + f (y ) = 0 , for 0 < α < 1, the results are similar to those
obtained above.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

References

[1] A. Biswas, Q. Zhou, M.Z. Ullah, H. Triki, S.P. Moshokoa, M. Belic, Optik 143 (2017) 131–134.
[2] A. Biswas, Quant. Phys. Lett. 1 (2) (2013) 79–83.
[3] H. Triki, A.M. Waz-waz, Rom. J. Phys. 61 (2016) 360–366.
[4] V.N. Serkin, T.L. Belyaeva, I.V. Alexandrov, G.M. Melchor, Proceedings Volume 4271, Optical Pulse and Beam Propagation III, (2001), https://doi.org/10.1117/
12.424706.
[5] A.V. Buryak, R.A. Sammut, B.A. Malomed, L.-C. Crasovan, D. Mihalache, Phys. Lett. A 288 (2001) 292–298.
[6] M. Naber, J. Math. Phys. 45 (2004) 3339.
[7] L. Wei, Y. He, X. Zhang, S. Wang, Finite Elem. Anal. Des. 59 (2012) 28–34.
[8] A. Esen, T.A. Sulaiman, H. Bulut, H.M. Baskonus, Optik 167 (2018) 150–156.
[9] R.K. Gazizov, A.A. Kasatkin, S. Yu Lukashchuk, Phys. Scr. T136 (2009) 014016.
[10] Y. Zhang, Y. Zhou, Nonlinear Dyn. 73 (2013) 783–793.
[11] A. Atangana, A.H. Cloot, Adv. Differ. Equ. (2013) 10.
[12] M. Inc, A. Yusuf, A.I. Aliyu, D. Baleanu, Physica A 496 (2018) 371–383.
[13] K. Singla, R.K. Gupta, J. Math. Phys. 57 (2016) 101504.
[14] N.H. Ibragimov, J. Phys. A: Math. Theor. 44 (2011) 432002.
[15] S.Y. Lukaschuk, Nonlinear Dyn. 80 (2015) 791–802.
[16] N.H. Ibragimov, A.H. Kara, F.M. Mahomed, Nonlinear Dyn. 15 (1998) 115–136.