Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
The Riccati Sub-ODE Method For FractionalDifferential-difference Equations
BIN ZHENGShandong University of Technology
School of ScienceZhangzhou Road 12, Zibo, 255049
Abstract: In this paper, we are concerned with seeking exact solutions for fractional differential-difference equa-tions by an extended Riccati sub-ODE method. The fractional derivative is defined in the sense of the modifiedRiemann-liouville derivative. By a combination of this method and a fractional complex transformation, the it-erative relations from indices n to n ± 1 are established. As for applications, we apply this method to solve thetwo-component fractional Volterra lattice equations and the fractional m-KdV lattice equation. Some new exactsolutions for the two fractional differential-difference equations are obtained.
Key–Words: Fractional differential-difference equations; Exact solutions; Riccati sub-ODE method; Fractionalcomplex transformations; Traveling wave solutions; Nonlinear evolution equationsMSC 2010: 35Q51; 35Q53
1 Introduction
Nonlinear differential equations (NLDEs) and nonlin-ear differential-difference equations (NLDDEs) canfind their applications in many aspects of mathemati-cal physics. In the last decades, research on seekingexact solutions for NLDEs and NLDDEs has been ahot topic, and many effective methods have been pre-sented so far (see [1-28] and the references therein).Among these investigations, we notice that little atten-tion is paid to fractional differential-difference equa-tions (FDDEs).
In this paper, we extend the Riccati sub-ODEmethod to seek exact solutions for FDDEs. The frac-tional derivative is defined in the sense of modifiedRiemann-liouville derivative [29-33] as follows.
Dγt f(t) =
1
Γ(1− γ)ddt
∫ t
0(t− ξ)−γ(f(ξ)− f(0))dξ,
0 < γ < 1,(f (n)(t))(γ−n), n ≤ γ < n+ 1, n ≥ 1.
Based on a fractional complex transformation, agiven fractional differential-difference equation canbe turned into another differential-difference equationof integer order, and the iterative relations of whichfrom indices n to n ± 1 are also established. By thisapproach, we will solve two fractional differential-difference equations: the two-component fractional
Volterra lattice equations{Dγt un = un(vn − vn−1),
Dγt vn = vn(un+1 − un),
(1)
and the following fractional m-KdV lattice equation
Dγt un = (α− u2n)(un+1 − un−1), (2)
where 0 < γ ≤ 1, un = un(t), vn = vn(t), n ∈Z, and Dγ
t denotes the modified Riemann-liouvillederivative of order γ with respect to the variable t.
When γ = 1, Eqs. (1) become the known two-component Volterra lattice equations [34], while Eq.(2) becomes the m-KdV lattice equation [34].
The following properties for the modifiedRiemann-Liouville are known to us (see [30-33]):
Dγt tr =
Γ(1 + r)
Γ(1 + r − γ)tr−γ , (3)
Dγt (f(t)g(t)) = g(t)Dγ
t f(t) + f(t)Dγt g(t), (4)
Dγt f [g(t)] = f ′g[g(t)]D
γt g(t) = Dγ
gf [g(t)](g′(t))γ .
(5)
WSEAS TRANSACTIONS on MATHEMATICS Bin Zheng
E-ISSN: 2224-2880 192 Volume 13, 2014
2 Description of the extended Ric-cati sub-ODE method for frac-tional differential-difference equa-tions
The main steps of the extended Riccati sub-ODEmethod for solving fractional differential-differenceequations are summarized as follows:
Step 1. Consider a fractional differential differ-ence equation in the form
P (un+p1(x), ..., un+pk(x), ..., Dγun+p1(x), ...,
Dγun+pk(x), D2γun+p1(x), ..., D
2γun+pk(x), ...) = 0,
(6)where the dependent variable u hasM components ui,the continuous variable x has N components xj , thediscrete variable n has Q components ni, the k shiftvectors ps ∈ ZQ has Q components psj , Dkγ denotesthe collection of mixed derivative terms of order kγ,and the order of the derivatives with respect to everyxj are integer multiples of γ.
Step 2. Using a fractional complex transfor-
mation Xj =xγj
Γ(1 + γ), and letting un+ps(x) =
Un+ps(X), Eq. (6) can be turned into
P (Un+p1(X), ..., Un+pk(X), ..., U ′n+p1(X), ...,
U ′n+pk
(X), U ′′n+p1(X), ..., U ′′
n+pk(X), ...) = 0, (7)
Step 2. Using a wave transformation
Un+ps(X) = Un+ps(ξn), ξn =
Q∑i=1
dini+
N∑j=1
cjXj+ζ,
where di, cj , ζ are all constants, we can rewrite Eq.(7) in the following form
˜P (Un+p1(ξn), ..., Un+pk(ξn), ..., U
′n+p1(ξn), ...,
U ′n+pk
(ξn), ..., U(r)n+p1(ξn), ..., U
(r)n+pk
(ξn)) = 0. (8)
Step 3: Suppose the solutions of Eq. (8) can bedenoted by
Un(ξn) =
l∑i=0
aiϕi(ξn), (9)
where ai are constants to be determined later, l is apositive integer that can be determined by balancingthe highest order linear term with the nonlinear termsin Eq. (8), ϕ(ξn) satisfies the known Riccati equation:
ϕ′(ξn) = σ + ϕ2(ξn). (10)
Step 4: We present some special solutionsϕ1, ..., ϕ6 for Eq. (10):
When σ < 0:
ϕ1(ξn) = −√−σ tanh(
√−σξn + c0),
ϕ2(ξn) = −√−σ coth(
√−σξn + c0),
ϕ1,2(ξn+ps) =
ϕ1,2(ξn)−√−σ tanh(
√−σ
Q∑i=1
dipsi)
1− ϕ1,2(ξn)√−σ
tanh(√−σ
Q∑i=1
dipsi)
,
(11)where c0 is an arbitrary constant.
When σ > 0:
ϕ3(ξn) =√σ tan(
√σξn + c0),
ϕ4(ξn) = −√σ cot(
√σξn + c0),
ϕ3,4(ξn+ps) =
ϕ3,4(ξn) +√σ tan(
√σ
Q∑i=1
dipsi)
1− ϕ3,4(ξn)√σ
tan(√σ
Q∑i=1
dipsi)
,
(12)and
ϕ5(ξn) =√σ[tan(2
√σξn + c0) + | sec(2
√σξn + c0)|],
ϕ5(ξn+ps) =
ϕ(1)5 (ξn) +
√σ tan(2
√σ
Q∑i=1
dipsi)
1− ϕ(1)5 (ξn)√σ
tan(2√σ
Q∑i=1
dipsi)
+
ϕ(2)5 (ξn) sec(2
√σ
Q∑i=1
dipsi)
1− ϕ(1)5 (ξn)√σ
tan(2√σ
Q∑i=1
dipsi)
,
(13)where ϕ(1)5 (ξn) =
√σtan(2
√σξn + c0), ϕ
(2)5 (ξn) =√
σ|sec(2√σξn+c0)|, and c0 is an arbitrary constant.
When σ = 0:
ϕ6(ξn) = − 1
ξn + c0,
ϕ6(ξn+ps) =ϕ6(ξn)
1− ϕ6(ξn)
Q∑i=1
dipsi
, (14)
where c0 is an arbitrary constant.Step 5: Substituting (9) into Eq. (8), by use of
Eqs. (10)-(14), the left hand side of Eq. (8) can be
WSEAS TRANSACTIONS on MATHEMATICS Bin Zheng
E-ISSN: 2224-2880 193 Volume 13, 2014
converted into a polynomial in ϕ(ξn). Equating eachcoefficient of ϕi(ξn) to zero, yields a set of algebraicequations. Solving these equations, we can obtain thevalues of ai, di, cj .
Step 6: Substituting the values of ai into (9), andcombining with the various solutions of Eq. (10), wecan obtain a variety of exact solutions for Eq. (6).
3 Application of the extended Ric-cati sub-ODE method to the two-component Volterra lattice equa-tions
In this section, we apply the extended Riccati sub-ODE method described in Section 2 to solve the two-component Volterra lattice equations denoted by Eqs.(1).
Using a fractional complex transformation T =tγ
Γ(1 + γ), and letting un+ps(t) = Un+ps(T ), ps =
0,±1, by (3) we have Dγt T = 1, and furthermore, by
the first equality in (5), Eqs. (1) can be turned into
{U ′n(T ) = Un(T )[Vn(T )− Vn−1(T )],
V ′n = Vn(T )[Un+1(T )− Un(T )],
(15)
Letting
Un(T ) = Un(ξn), Vn(T ) = Vn(ξn),
ξn = d1n+ c1T + ζ, (16)
where d1, c1, ζ are all constants, Eqs. (15) can berewritten as the following form:
{c1U
′n(ξn) = Un(ξn)[Vn(ξn)− Vn−1(ξn−1)],
c1V′n(ξn) = Vn(ξn)[Un+1(ξn+1)− Un(ξn)],
(17)Suppose the solutions for (17) can be denoted by
Un(ξn) =
l1∑i=0
aiϕi(ξn), (18)
Vn(ξn) =
l2∑i=0
biϕi(ξn), (19)
where ϕ(ξn) satisfies Eq. (10). In Eqs. (17), by bal-ancing the order of U ′
n and UnVn, the order of V ′n and
VnUn, we obtain l1 = l2 = 1. So we have
Un(ξn) = a0 + a1ϕ(ξn). (20)
Vn(ξn) = b0 + b1ϕ(ξn). (21)
We will proceed to solve Eqs. (17) in severalcases.
Case 1: If σ < 0, and assume (10) and (11) hold,then substituting (20), (21), (10) and (11) into Eqs.(17), collecting the coefficients of ϕi1,2(ξn) and equat-ing them to zero, we obtain a series of algebra equa-tions. Solving these equations, yields
a1 = −c1, a0 = − c1√−σ
tanh(√−σd1)
,
b1 = c1, b0 = − c1√−σ
tanh(√−σd1)
,
d1 = d1, c1 = c1,
or
a1 =a0 tanh(
√−σd1)√
−σ, a0 = a0,
b1 = −a0 tanh(√−σd1)√
−σ, b0 = a0,
c1 = −a0 tanh(√−σd1)√
−σ, d1 = d1.
So we obtain the following four groups of solitarywave solutions:
un(t) = − c1√−σ
tanh(√−σd1)
+c1√−σ tanh[
√−σ(d1n+ c1t
γ
Γ(1 + γ)+ ζ) + c0],
vn(t) = − c1√−σ
tanh(√−σd1)
−c1√−σ tanh[
√−σ(d1n+ c1t
γ
Γ(1 + γ)+ ζ) + c0],
(22)
un(t) = − c1√−σ
tanh(√−σd1)
+c1√−σ coth[
√−σ(d1n+ c1t
γ
Γ(1 + γ)+ ζ) + c0],
vn(t) = − c1√−σ
tanh(√−σd1)
−c1√−σ coth[
√−σ(d1n+ c1t
γ
Γ(1 + γ)+ ζ) + c0],
(23)where d1, c1, c0 are arbitrary constants, and
un(t) = a0 − a0 tanh(√−σd1)×
tanh[√−σ(d1n− a0 tanh(
√−σd1)tγ√
−σΓ(1 + γ)+ ζ) + c0],
vn(t) = a0 + a0 tanh(√−σd1)×
tanh[√−σ(d1n− a0 tanh(
√−σd1)tγ√
−σΓ(1 + γ)+ ζ) + c0],
(24)
WSEAS TRANSACTIONS on MATHEMATICS Bin Zheng
E-ISSN: 2224-2880 194 Volume 13, 2014
un(t) = a0 − a0 tanh(√−σd1)×
coth[√−σ(d1n− a0 tanh(
√−σd1)tγ√
−σΓ(1 + γ)+ ζ) + c0],
vn(t) = a0 + a0 tanh(√−σd1)×
coth[√−σ(d1n− a0 tanh(
√−σd1)tγ√
−σΓ(1 + γ)+ ζ) + c0],
(25)where d1, c0, a0 are arbitrary constants.
In Figs 1-2, the solitary wave solutions (22) withsome special parameters are demonstrated.
Case 2: If σ > 0, and assume (10) and (12) hold,then substituting (20), (21), (10) and (12) into Eqs.(17), collecting the coefficients of ϕi3,4(ξn) and equat-ing them to zero, we obtain a series of algebra equa-tions. Solving these equations, yields
a1 = −c1, a0 = − c1√σ
tan(√σd1)
, b1 = c1,
b0 = − c1√σ
tan(√σd1)
, d1 = d1, c1 = c1,
or
a1 =a0 tan(
√σd1)√
σ, a0 = a0, b1 = −a0 tan(
√σd1)√
σ,
b0 = a0, c1 = −a0 tan(√σd1)√
σ, d1 = d1.
So we obtain the following periodic wave solutions:
un(t) = − c1√σ
tan(√σd1)
−c1√σ tan[
√σ(d1n+ c1t
γ
Γ(1 + γ)+ ζ) + c0],
vn(t) = − c1√σ
tan(√σd1)
+c1√σ tan[
√σ(d1n+ c1t
γ
Γ(1 + γ)+ ζ) + c0],
(26)
un(t) = − c1√σ
tan(√σd1)
+c1√σ cot[
√σ(d1n+ c1t
γ
Γ(1 + γ)+ ζ) + c0],
vn(t) = − c1√σ
tan(√σd1)
−c1√σ cot[
√σ(d1n+ c1t
γ
Γ(1 + γ)+ ζ) + c0],
(27)where d1, c1, c0 are arbitrary constants, and
un(t) = a0 + a0 tan(√σd1)×
tan[√σ(d1n− a0 tan(
√σd1)t
γ√σΓ(1 + γ)
+ ζ) + c0],
vn(t) = a0 − a0 tan(√σd1)×
tan[√σ(d1n− a0 tan(
√σd1)t
γ√σΓ(1 + γ)
+ ζ) + c0],
(28)
WSEAS TRANSACTIONS on MATHEMATICS Bin Zheng
E-ISSN: 2224-2880 195 Volume 13, 2014
un(t) = a0 − a0 tan(√σd1)×
cot[√σ(d1n− a0 tan(
√σd1)t
γ√σΓ(1 + γ)
+ ζ) + c0],
vn(t) = a0 + a0 tan(√σd1)×
cot[√σ(d1n− a0 tan(
√σd1)t
γ√σΓ(1 + γ)
+ ζ) + c0],
(29)where d1, c0, a0 are arbitrary constants.
In Figs 3-4, the periodic wave solutions (26) withsome special parameters are demonstrated.
Case 3: If σ > 0, and assume (10) and (13) hold,then substituting (20), (21), (10) and (13) into Eqs.(17), using [ϕ
(2)5 (ξn)]
2 = σ + [ϕ(1)5 (ξn)]
2, collectingthe coefficients of [ϕ(1)5 (ξn)]
i[ϕ(2)5 (ξn)]
j and equatingthem to zero, we obtain a series of algebra equations.Solving these equations, we get that
a1 = −c1, a0 = 0, b1 = c1,
b0 = 0, d1 =π
2√σ, c1 = c1,
or
a1 = −c1, a0 = b0, b1 = c1,
b0 = b0, d1 =1
2√σarcsin(−2c1b0
√σ
b20 + c21σ), c1 = c1.
So we obtain the following trigonometric function so-
lutions:
un(t) = −c1√σ{tan[2
√σ( π
2√σn+ c1t
γ
Γ(1 + γ)+ ζ)
+c0] + | sec[2√σ( π
2√σn+ c1t
γ
Γ(1 + γ)+ ζ) + c0]|},
vn(t) = c1√σ{tan[2
√σ( π
2√σn+ c1t
γ
Γ(1 + γ)+ ζ)
+c0] + | sec[2√σ( π
2√σn+ c1t
γ
Γ(1 + γ)+ ζ) + c0]|},
(30)where c1, c0 are an arbitrary constants, and
un(t) = −c1√σ{tan[2
√σ( 1
2√σarcsin(−2c1b0
√σ
b20 + c21σ)n
+ c1tγ
Γ(1 + γ)+ ζ) + c0]
+| sec[2√σ( 1
2√σarcsin(−2c1b0
√σ
b20 + c21σ)n
+ c1tγ
Γ(1 + γ)+ ζ) + c0]|}+ b0,
vn(t) = c1√σ{tan[2
√σ( 1
2√σarcsin(−2c1b0
√σ
b20 + c21σ)n
+ c1tγ
Γ(1 + γ)+ ζ) + c0]
+| sec[2√σ( 1
2√σarcsin(−2c1b0
√σ
b20 + c21σ)n
+ c1tγ
Γ(1 + γ)+ ζ) + c0]|}+ b0,
(31)where c1, b0, c0 are an arbitrary constants.
Case 4: If σ = 0, and assume (10) and (14) hold,then substituting (20), (21), (10) and (14) into Eqs.(17), collecting the coefficients of ϕi6(ξn) and equat-ing them to zero, we obtain a series of algebra equa-tions. Solving these equations, yields
a1 = d1b0, a0 = b0, b1 = −d1b0,
b0 = b0, d1 = d1, c1 = −d1b0.
WSEAS TRANSACTIONS on MATHEMATICS Bin Zheng
E-ISSN: 2224-2880 196 Volume 13, 2014
Then we obtain the following rational solutions:un(t) =
−d1b0d1n− d1b0t
γ
Γ(1 + γ)+ ζ + c0
+ b0,
vn(t) =d1b0
d1n− d1b0tγ
Γ(1 + γ)+ ζ + c0
+ b0,(32)
where d1, b0, c0 are an arbitrary constants.
Remark 1 In [34, Eqs. (46), (47), (51), (52)], Ayhanand Bekir presented some exact solutions for the two-component Volterra lattice equations by the (G’/G)-expansion method. We note that our results (22), (23)are generalizations of [34, Eqs. (46), (47)], while(26), (27) are generalizations of [34, Eqs. (51), (52)].In fact, if we let
γ = 1, c0 = arth(C2
C1), σ =
4µ− λ2
4
or
γ = 1, c0 = arcoth(C1
C2), σ =
4µ− λ2
4,
then our results (22), (23) reduce to [34, Eq. (46),(47)]. If we let
γ = 1, c0 = arctan(−C2
C1), σ =
4µ− λ2
4
or
γ = 1, c0 = arccot(−C1
C2), σ =
4µ− λ2
4,
then our results (26), (27) reduce to [34, Eq. (51),(52)].
Remark 2 The established results by (30-32) are newexact solutions for the two-component Volterra latticeequations so far to our best knowledge.
4 Application of the extended Riccatisub-ODE method to the fractionalm-KdV lattice equation
In this section, we apply the extended Riccati sub-ODE method to solve the fractional m-KdV latticeequation denoted by Eq. (2).
Using a fractional complex transformation T =tγ
Γ(1 + γ), and letting un+ps(t) = Un+ps(T ), ps =
0,±1, by (3) we have Dγt T = 1, and furthermore, by
the first equality in (5), Eq. (2) can be turned into
U ′n(T ) = [α− U2
n(T )][Un+1(T )− Un−1(T )]. (33)
Letting
Un(T ) = Un(ξn), ξn = d1n+ c1T + ζ, (34)
where d1, c1, ζ are all constants, Eq. (33) can berewritten in the following form:
c1U′n(ξn)−(α−U2
n(ξn))(Un+1(ξn+1)−Un−1(ξn−1)) = 0.(35)
Suppose the solutions Un(ξn) for Eq. (3.3) can bedenoted by
Un(ξn) =
l∑i=0
aiϕi(ξn), (36)
where ϕ(ξn) satisfies Eq. (10). By balancing the high-est order linear term with the nonlinear terms in Eq.(35) we obtain l+1 = 2l, and then l = 1. So we have
Un(ξn) = a0 + a1ϕ(ξn). (37)
We will proceed to solve Eq. (35) in several cases.
Case 1: If σ < 0, and assume (10) and (11) hold,then substituting (37), (10) and (11) into Eq. (35),collecting the coefficients of ϕi1,2(ξn) and equatingthem to zero, we obtain a series of algebraic equa-tions. Solving these equations, yields
a1 = ±√
−ασtanh(
√−σd1), a0 = 0,
d1 = d1, c1 =2α√−σ
tanh(√−σd1).
So we obtain the following solitary wave solutions:
un(t) = ±√α tanh(
√−σd1)
tanh[√−σ(d1n+
2α√−σ
tanh(√−σd1)
tγ
Γ(1 + γ)+ζ)+c0],
(38)and
un(t) = ±√α tanh(
√−σd1)
coth[√−σ(d1n+
2α√−σ
tanh(√−σd1)
tγ
Γ(1 + γ)+ζ)+c0],
(39)
WSEAS TRANSACTIONS on MATHEMATICS Bin Zheng
E-ISSN: 2224-2880 197 Volume 13, 2014
where d1, c0 are arbitrary constants.Case 2: If σ > 0, and assume (10) and (12)
hold, then substituting (37), (10) and (12) into Eq.(35), collecting the coefficients of ϕi3,4(ξn) and equat-ing them to zero, we obtain a series of algebraic equa-tions. Solving these equations, yields
a1 = ±√α
σtan(
√σd1), a0 = 0,
d1 = d1, c1 =2α√σtan(
√σd1).
Then we have the following periodic wave solutions:
un(t) = ±√α tan(
√σd1)
tan[√σ(d1n+
2α√σtan(
√σd1)
tγ
Γ(1 + γ)+ζ)+c0], (40)
andun(t) = ±
√α tan(
√σd1)
cot[√σ(d1n+
2α√σtan(
√σd1)
tγ
Γ(1 + γ)+ζ)+c0], (41)
where d1, c0 are arbitrary constants.In [34, Eqs. (32) and (36)], Ayhan and Bekir pre-
sented some exact solutions for m-KdV lattice equa-tion by the (G’/G)-expansion method as follows:
un = ±√α tanh(
√λ2 − 4µ
2d1)×
(C1 sinh(
√λ2 − 4µ
2ξn) + C2 cosh(
√λ2 − 4µ
2ξn)
C1 cosh(
√λ2 − 4µ
2ξn) + C2 sinh(
√λ2 − 4µ
2ξn))
),
(42)
where ξn = d1n+4α√λ2 − 4µ
tanh(
√λ2−4µ2 d1)t+ζ,
and
un = ±√α tan(
√4µ− λ2
2d1)×
(−C1 sin(
√4µ− λ2
2ξn) + C2 cos(
√4µ− λ2
2ξn)
C1 cos(
√4µ− λ2
2ξn) + C2 sin(
√4µ− λ2
2ξn))
),
(43)
where ξn = d1n+ 4α√4µ− λ2
tan(
√4µ−λ22 d1)t+ ζ.
We note that our results (38) and (40) are solu-tions of more general forms than Eqs. (42) and (43).In fact, if we let γ = 1, c0 = arth(C2
C1), σ =
4µ− λ2
4 or γ = 1, c0 = arcoth(C1C2
), σ =4µ− λ2
4 ,then our result (38) reduces to (42). If we let γ =
1, c0 = arctan(−C2C1
), σ =4µ− λ2
4 or γ = 1, c0 =
arccot(−C1C2
), σ =4µ− λ2
4 , then our result (40) re-duces to (43).
Case 3: If σ > 0, and assume (10) and (13) hold,then substituting (37), (10) and (13) into Eq. (35), us-ing [ϕ
(2)5 (ξn)]
2 = σ+ [ϕ(1)5 (ξn)]
2, collecting the coef-ficients of [ϕ(1)5 (ξn)]
i[ϕ(2)5 (ξn)]
j and equating them tozero, we obtain a series of algebraic equations. Solv-ing these equations, we get three families of values asfollows:
a1 = ±√α
σ, a0 = 0,
d1 = − π
4√σ, c1 = − 2α√
σ,
a1 = ±√α
σ, a0 = 0,
d1 =π
4√σ, c1 =
2α√σ,
or
a1 = ±
√2α− α sin2(2
√σd1)− 2α cos(2
√σd1)
√σ sin(2
√σd1)
,
d1 = d1, a0 = 0, c1 = −2α(cos(2√σd1)− 1)√
σ sin(2√σd1)
So we obtain the following trigonometric function so-lutions:
un(t) = ±√α{tan[2
√σ(
π
4√σn+
2αtγ√σΓ(1 + γ)
+ ζ)
+c0]+ | sec[2√σ(
π
4√σn+
2αtγ√σΓ(1 + γ)
+ζ)+c0]|},
(44)
un(t) = ±√α{tan[2
√σ(− π
4√σn− 2αtγ√
σΓ(1 + γ)+ ζ)
+c0]+| sec[2√σ(− π
4√σn− 2αtγ√
σΓ(1 + γ)+ζ)+c0]|},
(45)where c0 is an arbitrary constant, and
WSEAS TRANSACTIONS on MATHEMATICS Bin Zheng
E-ISSN: 2224-2880 198 Volume 13, 2014
un(t) = ±
√2α− α sin2(2
√σd1)− 2α cos(2
√σd1)
sin(2√σd1)
{tan[2√σ(d1n− 2α(cos(2
√σd1)− 1)tγ√
σ sin(2√σd1)Γ(1 + γ)
+ ζ) + c0]
+| sec[2√σ(d1n−
2α(cos(2√σd1)− 1)tγ√
σ sin(2√σd1)Γ(1 + γ)
+ ζ)+ c0]|},
(46)where d1, c0 are arbitrary constants.
Case 4: If σ = 0, and assume (10) and (14) hold,then substituting (37), (10) and (14) into Eq. (35), col-lecting the coefficients of ϕi6(ξn) and equating them tozero, we obtain a series of algebraic equations. Solv-ing these equations, yields
a1 = ±√αd1, a0 = 0, d1 = d1, c1 = 2d1α.
Then we obtain the following rational solution:
un(t) =±√αd1
d1n+2d1αt
γ
Γ(1 + γ)+ ζ + c0
, (47)
where d1, c0 are arbitrary constants.
Remark 3 Our results (44)-(47) have not been re-ported by other authors so far to our best knowledge.
5 Conclusions
The Riccati sub-ODE method is extended to seekexact solutions for fractional differential-differenceequations. By this method, we solved the two-component fractional Volterra lattice equations andthe fractional m-KdV lattice equation successfully,and as a result, some generalized exact solutions in-cluding solitary wave solutions, periodic wave solu-tions and rational function solutions for them havebeen found with the aid of mathematical software. Be-ing concise and effective, we note that this approachcan be applied to solve other fractional differential-difference equations.
References:
[1] M. L. Wang and X. Z. Li, Applications of F-expansion to periodic wave solutions for a newHamiltonian amplitude equation, Chaos, Soli-tons and Fractals, 24, 2005, pp. 1257-1268.
[2] J. H. He and X. H. Wu, Exp-function method fornonlinear wave equations, Chaos, Solitons andFractals, 30, 2006, pp. 700-708.
[3] M. L. Wang, X. Z. Li and J. L. Zhang, The(G’/G)-expansion method and travelling wavesolutions of nonlinear evolution equations inmathematical physics, Phys. Lett. A, 372, 2008,pp. 417-423.
[4] E. M. E. Zayed and K. A. Gepreel, The mod-ified (G’/G)-expansion method and its applica-tions to construct exact solutions for nonlinearPDEs, WSEAS Transactions on Mathematics,10(8), 2011, pp. 270-278.
[5] E. M. E. Zayed, A further improved (G’/G)-expansion method and the extended tanh-method for finding exact solutions of nonlin-ear PDEs, WSEAS Transactions on Mathemat-ics, 10(2), 2011, pp. 56-64.
[6] E. M. E. Zayed and M. Abdelaziz, Exact trav-eling wave solutions of nonlinear variable co-efficients evolution equations with forced termsusing the generalized (G’/G)-expansion method,WSEAS Transactions on Mathematics, 10(3),2011, pp. 115-124.
[7] Q. H. Feng and B. Zheng, Traveling WaveSolutions for the Fifth-Order Sawada-KoteraEquation and the General Gardner Equationby (G’/G)-Expansion Method, WSEAS Transac-tions on Mathematics, 9(3), 2010, pp. 171-180.
[8] Y. Wang, Variable-coefficient Simplest EquationMethod For Solving Nonlinear Evolution Equa-tions In Mathematical Physics, WSEAS Transac-tions on Mathematics, 12(5), 2013, pp. 512-520.
[9] E. M. E. Zayed and S. Al-Joudi, An Improved(G’/G)-expansion Method for Solving NonlinearPDEs in Mathematical Physics, ICNAAM 2010,AIP. Conf. Proc., Vol. 1281, 2010, pp. 2220-2224.
[10] Q. H. Feng and B. Zheng, Traveling WaveSolutions for the Fifth-Order Kdv Equationand the BBM Equation by (G’/G)-ExpansionMethod, WSEAS Transactions on Mathematics,9(3), 2010, pp. 201-210.
[11] B. Zheng, Application Of A GeneralizedBernoulli Sub-ODE Method For Finding Trav-eling Solutions Of Some Nonlinear Equations,WSEAS Transactions on Mathematics, 11(7),2012, pp. 634-642.
[12] C. C. Kong, D. Wang, L. N. Song andH. Q. Zhang, New exact solutions to MKDV-Burgers equation and (2+1)-dimensional disper-sive long wave equation via extended Riccatiequation method, Chaos, Solitons and Fractals,39, 2009, pp. 697-706.
WSEAS TRANSACTIONS on MATHEMATICS Bin Zheng
E-ISSN: 2224-2880 199 Volume 13, 2014
[13] M. Eslami, A. Neyrame and M. Ebrahimi, Ex-plicit solutions of nonlinear (2+1)-dimensionaldispersive long wave equation, J. King SaudUniver.-Sci. , 24, 2012, pp. 69-71.
[14] I. Aslan, Discrete exact solutions to some non-linear differential-difference equations via the(G’/G)-expansion method, Appl. Math. Comput.,215, 2009, pp. 3140-3147.
[15] X. B. Hu and W. X. Ma, Application of Hirota’sbilinear formalism to the Toeplitz lattice somespecial soliton-like solutions, Phys. Lett. A, 293,2002, pp. 161-165.
[16] B. Tang, Y. N. He, L. L. Wei and S. L. Wang,Variable-coefficient discrete (G’/G)-expansionmethod for nonlinear differential-differenceequations, Phys. Lett. A , 375, 2011, pp. 3355-3361.
[17] W. Zhen, Discrete tanh method for non-linear difference-differential equations, Com-put. Phys. Commun., 180, 2009, pp. 1104-1108.
[18] I. Aslan, A discrete generalization of theextended simplest equation method, Com-mun. Nonlinear Sci. Numer. Simul., 15, 2010,pp. 1967-1973.
[19] C. Q. Dai, X. Cen and S. S. Wu, Exact travel-ling wave solutions of the discrete sine-Gordonequation obtained via the exp-function method,Nonlinear Anal., 70, 2009, pp. 58-63.
[20] C. S. Liu, Exponential function rational expan-sion method for nonlinear differential-differenceequations, Chaos, Solitons and Fractals, 40,2009, pp. 708-716.
[21] S. Zhang, L. Dong, J. M. Ba and Y. N. Sun,The (G’/G)-expansion method for nonlineardifferential-difference equations, Phys. Lett. A,373, 2009, pp. 905-910.
[22] B. Ayhan and A. Bekir, The (G’/G)-expansionmethod for the nonlinear lattice equations, Com-mun. Nonlinear Sci. Numer. Simulat., 17, 2012,pp. 3490-3498.
[23] H. Xin, The exponential function rational ex-pansion method and exact solutions to nonlinearlattice equations system, Appl. Math. Comput.,217, 2010, pp. 1561-1565.
[24] K. A. Gepreel and A. R. Shehata, Rational Ja-cobi elliptic solutions for nonlinear differential-difference lattice equations, Appl. Math. Lett.,25, 2012, pp. 1173-1178.
[25] W. H. Huang and Y. L. Liu, Jacobi elliptic func-tion solutions of the Ablowitz-Ladik discretenonlinear Schrodinger system, Chaos, Solitonsand Fractals, 40, 2009, pp. 786-792.
[26] C. B. Wen, Traveling Wave Solutions For TwoNonlinear Lattice Equations By An ExtendedRiccati Sub-equation Method, WSEAS Transac-tions on Mathematics, 11(12), 2012, pp. 1085-1093.
[27] C. B. Wen and B. Zheng, A New Fractional Sub-equation Method For Fractional Partial Differen-tial Equations, WSEAS Transactions on Mathe-matics, 12(5), 2013, pp. 564-571.
[28] Q. H. Feng and B. Zheng, Traveling Wave So-lutions For The Variant Boussinseq EquationAnd The (2+1)-dimensional Nizhnik-Novikov-Veselov (NNV) System By (G’/G)-expansionmethod, WSEAS Transactions on Mathematics,9(3), 2010, pp. 191-200.
[29] G. Jumarie, Modified Riemann-Liouville deriva-tive and fractional Taylor series of non-differentiable functions further results, Com-put. Math. Appl., 51, 2006, pp. 1367-1376.
[30] B. Lu, Backlund transformation of fractionalRiccati equation and its applications to non-linear fractional partial differential equations,Phys. Lett. A, 376, 2012, pp. 2045-2048.
[31] S. M. Guo, L. Q. Mei, Y. Li and Y. F. Sun,The improved fractional sub-equation methodand its applications to the space-time fractionaldifferential equations in fluid mechanics, Phys.Lett. A, 376, 2012, pp. 407-411.
[32] S. Zhang and H. Q. Zhang, Fractional sub-equation method and its applications to nonlin-ear fractional PDEs, Phys. Lett. A, 375, 2011,pp. 1069-1073.
[33] B. Zheng, (G’/G)-Expansion Method for Solv-ing Fractional Partial Differential Equations inthe Theory of Mathematical Physics, Com-mun. Theor. Phys. (Beijing, China), 58, 2012,pp. 623-630.
[34] B. Ayhan and A. Bekir, The (G’/G)-expansionmethod for the nonlinear lattice equations, Com-mun. Nonlinear Sci. Numer. Simulat., 17, 2012,pp. 3490-3498.
WSEAS TRANSACTIONS on MATHEMATICS Bin Zheng
E-ISSN: 2224-2880 200 Volume 13, 2014