The fractional partial differential equations stand for natural phenomena all over the world from science to engineering. When it comes to obtaining the solutions of these equations, there are many various techniques in the literature. Some of these give to us approximate solutions; others give to us analytical solutions. In this paper, we applied the modified trial equation method (MTEM) to the one-dimensional nonlinear fractional wave equation (FWE) and time fractional generalized Burgers equation. Then, we submitted 3D graphics for different value of α.
1. Introduction
All over the world, a physical event may depend not only on the time but also on the previous process, which can be successfully formed by using the theory of derivatives and integrals of fractional order. These processes represent different physical problems in the manner of variable order. In this sense, the fractional differential equations have been used for the definition of nonlinear phenomena in applied science, physics, chemistry, engineering, and other areas of science. In order to solve these problems, a general method cannot be defined even in the most useful works. Also, a remarkable progress has been become in the construction of the approximate solutions for fractional nonlinear partial differential equations [1–3]. Several powerful methods [4–13] have been proposed to obtain approximate and exact solutions of fractional differential equations, such as the Sumudu transform method, the Homotopy analysis method, and the homotopy perturbation method.
Liu introduced a new approach called the complete discrimination system for a polynomial to classify the traveling wave solutions as nonlinear evolution equations and applied this idea to some nonlinear partial differential equations [14, 15]. So, to the best of our knowledge, the modified trial equation method has not been widely applied for studying the invariance properties of fractional PDEs. Furthermore, some authors [16, 17] used the trial equation method proposed by Liu. However, we established a new modified trial equation method to obtain 1-soliton, singular soliton, hyperbolic function solutions [18, 19], elliptic integral function and Jacobi elliptic function solutions, or the others to nonlinear partial differential equations with generalized evolution in [20–22].
In Section 2, primarily we give some definitions and properties of the fractional calculus and also produce a new modified trial equation method for fractional nonlinear evolution equations with higher order nonlinearity. The power of this steerable method showed that this method can be applied to different equations. In Section 3, as an application, we solve the nonlinear fractional partial differential equation such as one-dimensional nonlinear fractional wave equation [23]:
(1)∂αu∂tα+auxx+βu+γu3=0,
where a, β, and γ are arbitrary constants and α is a parameter describing the order of the fractional time derivative. We consider time fractional generalized Burgers equation [27] described as follows:
(2)∂αu∂tα-uxx-βupux=0,
where 0<α≤1, p>0 which occur in different areas in mathematical physics; here the time fractional derivative leads to subdiffusion and subdispersion, respectively, and extend the Lie symmetry analysis to derive their infinitesimals [24]. In this research, we obtain the classification of the wave solutions to (1) and (2) and derive some new solutions. Using the modified trial equation method, we found some new exact solutions of the fractional nonlinear physical problem. The purpose of this paper is to obtain exact solutions of the one-dimensional nonlinear fractional wave equation by modified trial equation method.
2. Preliminaries
In this part of the paper, it would be helpful to give some definitions and properties of the fractional calculus theory. Here, we shortly review the modified Riemann-Liouville derivative from the recent fractional calculus proposed by Jumarie [25, 26]. Let f:[0,1]→ℜ be a continuous function and α∈(0,1). The Jumarie modified fractional derivative of order α and f may be defined by the expression of the following [23]:
(3)Dxαf(x)={1Γ(-α)∫0x(x-ξ)-α-1[f(ξ)-f(0)]dξ,hhhhhhhhhhhhhhhhhhhhhhhhhhhhα<0,1Γ(1-α)ddx×∫0x(x-ξ)-α[f(ξ)-f(0)]dξ,hhhhhhhhhhhhhhhhhhhhhhhh0<α<1,(f(n)(ξ))α-n,n≤α≤n+1,n≥1.
In addition to this expression, we may give a summary of the fractional modified Riemann-Liouville derivative properties which are used further in this paper. Some of the useful formulas are given as [23]
(4)Dxαk=0,Dxαxμ={0,μ≤α-1,Γ(μ+1)Γ(μ-α+1)xμ-x,μ>α-1.
In this paper, a new approach to the trial equation method will be given. In order to apply this method to fractional nonlinear partial differential equations, we consider the following steps.
Step 1.
We consider time fractional partial differential equation in two variables and a dependent variable u:
(5)P(u,Dtαu,ux,uxx,uxxx,…)=0,
and take the wave transformation
(6)u(x,t)=u(η),η=kx-λtαΓ(1+α),
where λ≠0. Substituting (6) into (5) yields a nonlinear ordinary differential equation:
(7)N(u,u′,u′′,u′′′,…)=0.
Step 2.
Take trial equation as follows:
(8)u′=F(u)G(u)=∑i=0naiui∑j=0lbjuj=a0+a1u+a2u2+⋯+anunb0+b1u+b2u2+⋯+blul,(9)u′′=F(u)(F′(u)G(u)-F(u)G′(u))G3(u),
where F(u) and G(u) are polynomials. Substituting above relations into (7) yields an equation of polynomial Ω(u) of u:
(10)Ω(u)=ρsus+⋯+ρ1u+ρ0=0.
According to the balance principle, we can get a relation of n and l. We can compute some values of n and l.
Step 3.
Letting the coefficients of Ω(u) all be zero will yield an algebraic equations system:
(11)ρi=0,i=0,…,s.
By solving this system, we will specify the values of a0,…,an and b0,…,bl.
Step 4.
Reduce (8) to the elementary integral form
(12)±(μ-μ0)=∫G(u)F(u)du.
Using a complete discrimination system for polynomial to classify the roots of F(u), we solve (12) with the help of Mathematica7 and classify the exact solutions to (7). In addition, we can write the exact traveling wave solutions to (5), respectively. For a better interpretation of results obtained in this way, we plotted 3D surfaces of (27), (40), (52), and (60) in Figures 1, 2, 3, and 4 by taking into consideration suitable parameter.
Graphics of the solution equation (27) corresponding to the values α=β=0.01, α=β=0.25, and α=β=0.75 from left to right when k=a1=1, a2=0.1, γ=-1, -5<x<5, and 0<t<5.
Graphics of the solution equation (40) corresponding to the values α=β=0.01, α=β=0.25, and α=β=0.75 from left to right when k=a1=b0=b1=γ=0.01, -5<x<5, and 0<t<5.
Graphics of the solution equation (52) corresponding to the values α=β=0.01, α=β=0.25, and α=β=0.75 from left to right when k=a2=b0=0.1, -5<x<5, and 0<t<5.
Graphics of the solution equation (60) corresponding to the values α=β=0.01, α=β=0.25, and α=β=0.75 from left to right when a1=b0=b1=p=k=1, -5<x<1, and 0<t<5.
3. Applications
In this section, we applied the method to the one-dimensional nonlinear fractional wave equation and time fractional generalized Burgers equation.
Example 1.
Firstly, we consider one-dimensional nonlinear fractional wave equation [23]. In the case of α=1, (1) reduces to the classical nonlinear one-dimensional nonlinear wave equation. Many researchers have tried to get the exact solutions of this equation by using a variety of methods.
Let us consider the travelling wave solutions of (1), and then and we perform the transformation u(x,t)=u(η) and η=kx-λtα/Γ(1+α) where k and λ are constants. Then, integrating this equation with respect to η and setting the integration constant to zero. When it comes to converting fractional order differential equation into differential equation with integer order, we can perform the following:
(13)dαudtα=dαudηαdαηdtα=-λu′,d2udx2=ddη(dudη)dηdx+dudηddη(dηdx)=k3u′′,
so, when we use dαu/dtα and d2u/dx2 in (1), we get ordinary differential equation as follows:
(14)-λu′+ak3u′′+βu+γu3=0.
When we rearrange to (8) and (9) for balance principle, we obtain the following:
(15)u′=anblun-l+⋯,(16)u′′=(annanbl-bllan2)bl3u2n-2l-1+⋯.
Balancing the highest order nonlinear terms of u′′ and u3 in (14), we get balance term for suitability
(17)2n-2l-1=3⟹n=l+2.
This resolution procedure is applied, and we obtain results as follows.
Case 1.
If we take l=0 and n=2, then
(18)u′=F(u)G(u)=∑i=0naiui∑j=0lbjuj=a0+a1u+a2u2b0,
and then
(19)u′′=F(u)[F′(u)G(u)-F(u)G′(u)]G3(u),=(a1+2a2u)(a0+a1u+a2u2)b02,
where a2≠0 and b0≠0 When we use u′ and u′′ in (14), we get a system of algebraic equations for (14). Thus, we have a system of algebraic equations from the coefficients of the polynomial of u. Solving the algebraic equation system (14) by using Mathematica programming yields the following coefficients:
(20)a0=0,a1=βb0k2a,a2=±-γb0k2a,b0=b0,λ=3kβa2.
By substituting these coefficients into (12), we have
(21)±(μ-μ0)=k2a-γ∫duu2+(a1/a2)u.
Integrating (21) by using Mathematica programming, we obtain the solutions to (1), as follows, for different values of the roots of the polynomial equation:
(22)±(μ-μ0)=-Au-α1,α1=α2,±(μ-μ0)=Aα1-α2ln|u-α1u-α2|,α1≠α2,
where A=±k2a/-γ, and also α1 and α2 are the roots of the polynomial equation as follows:
(23)u2+a1a2u=0.
Therefore, we find solutions
(24)u(x,t)=α1+A±(kx-(3kβatα/2Γ(1+α))-η0),(25)u(x,t)=α1±α1-α2exp[((α1-α2)/A)(kx-(3kβatα/2Γ(1+α))-η0)]-1.
For simplicity, if we take η0=0, then the solutions equations (24) and (25) can reduce to rational and single kink solution, respectively,
(26)u(x,t)=α1+AB1(x-λ1tα),(27)u(x,t)=α1±α1-α2exp[B2(x-λ1tα)]-1,
where B1=±k, B2=k(α1-α2)/A, and λ1=3βa/2Γ(1+α). Here, B1 and B2 are the inverse width of the soliton. We can regulate (27) to rewrite in the hyperbolic form as follows:
(28)u(x,t)=α1+α1-α2exp[B2(x-λ1tα)]-1,(29)u(x,t)=α2+α2-α1exp[B2(x-λ1tα)]-1.
If we consider the following equation for simplicity of (28):
(30)u(μ)=α1+α1-α2exp[μ]-1,
then, we get
(31)u(μ)=α1+α1-α2exp[μ]-1=α1exp[μ]-α2exp[μ]-1=α1exp[μ]-α2/α1exp[μ]-1.
If it takes α1=-α2 for (28), we get the hyperbolic function solution of (28):
(32)u(μ)=α1coth[μ2],
where μ=B2(x-λ1tα).
Remark 2.
The solutions equations (26)-(27) obtained by using the extended trial equation method for (1) have been checked by Mathematica. To our knowledge, the rational function solution and single kink solution that we found in this paper are not shown in the previous literature. These results are new traveling wave solutions of (1).
Case 2.
In the same way as in Case 1, If we take l=1 and n=3, then
(33)u′=a0+a1u+a2u2+a3u3b0+b1u,u′′=((a0+a1u+a2u2+a3u3)hhhh×((b0+b1u)(a1+2a2u+3a3u2)fffffffffffff-b1(a0+a1u+a2u2+a3u3)))×((b0+b1u)3)-1,
where a3≠0, b1≠0. Respectively, solving the algebraic equation system (11) yields the following:
(34)a0=0,a1=±βb0k2a,a2=-b0(-γ)-b1βk2a,a3=-b1γk2a,b0=b0,b1=b1,λ=±3kβa2.
Substituting these coefficients into (12), we have
(35)±(μ-μ0)=-γk2a∫u+b0/b1u3+(a2/a3)u2+(a1/a3)udu.
Integrating (35), we procure the solution to (1) as follows:
(36)±(μ-μ0)=-A1(b0+2b1u-b1α1)2b1(u-α1)2,(37)±(μ-μ0)=A1b1(b0+b1α2(α1-α2)2ln|u-α2u-α1|hhhhhh-b0+b1α1(u-α1)(α1-α2)),±(μ-μ0)=A1b1(α1-α2)(α2-α3)(α1-α3)×(ln|(u-α1)M(u-α3)N(u-α2)P|),
where A1=-γ/k2a, M=(α2-α3)(b0+b1α1), N=(α1-α2)(b0+b1α3) and P=(α1-α3)(b0+b1α2). Also α1, α2, and α3 are the roots of the polynomial equation
(38)u3+a2a3u2+a1a3u+a0a3=0.
Therefore, we find a solution from (36):
(39)u(x,t)=(A1b1(A1b1-2(b0+b1α1)(kx±3ktαβa2Γ(1+α)-η0))b1(2α1(kx±3ktαβa2Γ(1+α)-η0)-A1)+A1b1(A1b1-2(b0+b1α1)(kx±3ktαβa2Γ(1+α)-η0)))×(2b1(kx±3ktαβa2Γ(1+α)-η0))-1.
For simplicity, if we take η0=0, then the solution equation (39) can reduce to rational solution
(40)u(x,t)=(A1b1(A1b1-2(b0+b1α1)(kx±kλ1tα))b1(2α1(kx±kλ1tα)-A1)hhhh+A1b1(A1b1-2(b0+b1α1)(kx±kλ1tα)))×(2b1(kx±kλ1tα))-1.
Remark 3.
The solution equation (40) computed in Case 2 has been checked by Mathematica. We think that these solutions have not been found in the literature, and these results are new traveling wave solutions of (1).
Example 4.
Secondly, we consider the time fractional generalized Burgers equation [24] as follows:
(41)∂αu∂tα-uxx-βupux=0.
In the case of α=1 and p=1, (41) reduces to the well-known classical nonlinear Burgers equation. Many researchers have tried to get the exact solutions of this equation by using a different method [28, 29].
Let us consider the travelling wave solutions of (41), and then we perform the transformation u(x,t)=u(η) and η=kx-λtα/Γ(1+α) where k, λ are constants. Then, integrating this equation with respect to η and setting the integration constant to zero, we get
(42)-λu′(η)-k2u′′(η)-βkup(η)ux(η)=0.
When we conduct once more transformation
(43)u(η)=v1/p(η),
we get the following:
(44)λp(p+1)v-k2(p+1)v′-βkv2=0.
Substituting (8) into (44) and using balance principle yield the following:
(45)n=l+2.
This resolution procedure is applied, and we obtain results as follows.
Case 1.
If we take l=0 and n=2, then
(46)v′=a0+a1v+a2v2b0,
where a2≠0 and b0≠0. Thus, we have a system of algebraic equations from the coefficients of the polynomial of v. Solving the algebraic equation system (11) yields the following:
(47)a0=0,a1=a1,a2=-pβb0k+kp,b0=b0,λ=-k2a1pb0.
By substituting these coefficients into (11), we have
(48)±(μ-μ0)=-k+kppβ∫dvv2+(a1/a2)v.
By integrating (48), we procure the solution to (41) as follows:
(49)±(μ-μ0)=-A2v-α1,±(μ-μ0)=A2α1-α2ln|v-α1v-α2|,
where A2=-(k+kp)/pβ. By substituting the solutions equation (49) into (43), we found solutions of the following exact traveling wave solutions, such as rational function solution and single kink solution:
(50)u(x,t)=[α1±A2(kx+(ka1tα/pb0Γ(1+α))-η0)]1/p,u(x,t)=[α1±α1-α2exp[((α1-α2)/A2)(kx+(k2a1tα/pb0Γ(1+α))-η0)]-1]1/p.
For simplicity, if we take η0=0, then the solutions equation (50) can reduce to the following:
(51)u(x,t)=[α1±B3(x-λ2tα)]1/p,(52)u(x,t)=[α1±α1-α2exp[B4(x-λ2tα)]-1]1/p,
where B3=A2/k, B4=k(α1-α2)/A2, and λ2=-ka1/pb0Γ(1+α).
Remark 5.
The solutions equations (51) and (52) obtained by using the modified trial equation method for (41) have been checked by Mathematica. To our knowledge, the rational function solution and single kink solution that we found in this paper are new traveling wave solutions of (41).
Case 1.
If we take l=1 and n=3, then
(53)v′=a0+a1v+a2v2+a3v3b0+b1v,
where a3≠0, b1≠0. Thus, we have a system of algebraic equations from the coefficients of the polynomial of v. Solving the algebraic equation system (11) yields the following:
(54)a0=0,a1=a1,a2=a2,a3=-pβb0(k(1+p)a2+pβb0)k2(1+p)2a1,b0=b0,b1=-b0(k(1+p)a2+pβb0)k(1+p)a1,λ=-k2a1pb0.
By substituting these coefficients into (11), we have
(55)±(μ-μ0)=k(1+p)pβ∫v+b0/b1v3+(a2/a3)v2+(a1/a3)vdv,
where A3=(k+kp)/pβ. By integrating (55), we procure the solution to (41) as follows:
(56)±(μ-μ0)=-A3(b0+2b1v-b1α1)2b1(v-α1)2,(57)±(μ-μ0)=A3b1(b0+b1α2(α1-α2)2ln|v-α2v-α1|hhhhhhh-b0+b1α1(v-α1)(α1-α2)),±(μ-μ0)=A3b1(α1-α2)(α2-α3)(α1-α3)×(ln|(v-α1)M(v-α3)N(v-α2)P|),
where A3=k(1+p)/pβ, M=(α2-α3)(b0+b1α1), N=(α1-α2)(b0+b1α3), and P=(α1-α3)(b0+b1α2). Also α1, α2, and α3 are the roots of the polynomial equation:
(58)v3+a2a3v2+a1a3v+a0a3=0.
By substituting the solution equation (56) into (43), we found solution of the following exact traveling wave solutions, such as rational function solution:
(59)u(x,t)=[(b1(2α1(kx+k2a1tαpb0Γ(1+α)-η0)-A3)hhh+A3b1(A3b1-2(b0+b1α1)(kx+k2a1tαpb0Γ(1+α)-η0)))×(2b1(kx+k2a1tαpb0Γ(1+α)-η0))-1]1/p.
For simplicity, if we take η0=0, then the solution equation (59) can reduce to rational solution:
(60)u(x,t)=[A3b1(A3b1-2(b0+b1α1)(kx±kλ2tα))(A3b1(A3b1-2(b0+b1α1)(kx±kλ2tα))b1(2α1(kx±kλ2tα)-A3)hhh+A3b1(A3b1-2(b0+b1α1)(kx±kλ2tα)))hhh×(2b1(kx±kλ2tα))-1]1/p.
Remark 6.
The solution equation (60) computed in Case 2 has been checked by Mathematica. We think that these solutions have not been found in the literature, and these results are new traveling wave solutions of (41).
4. Conclusions
In this paper, the modified trial equation method has been applied to the one-dimensional nonlinear fractional wave equation and time fractional generalized Burgers equation. We used it to obtain some soliton and rational function solutions to the one-dimensional nonlinear fractional wave equation and time fractional generalized Burgers equation. This method is reliable and effective and gives several new solution functions such as rational function solutions and single kink solutions. We think that the proposed method can also be applied to other generalized fractional nonlinear differential equations. In our future studies, we will solve nonlinear fractional partial differential equations by this approach.
MillerK. S.RossB.1993New York, NY, USAJohn Wiley & Sonsxvi+366MR1219954KilbasA. A.SrivastavaH. M.TrujilloJ. J.2006204Amsterdam, The NetherlandsElsevier Science B.V.xvi+523North-Holland Mathematics StudiesMR221807310.1016/S0304-0208(06)80001-0PodlubnyI.1999198San Diego, Calif, USAAcademic Pressxxiv+340Mathematics in Science and EngineeringMR1658022BulutH.BaskonusH. M.TuluceS.The solution of wave equations by Sumudu transform method201243667210.5373/jaram.1317.021812MR2955863BulutH.BaskonusH. M.TuluceS.Homotopy perturbation Sumudu transform for heat equations2013414960SafariM.GanjiD. D.MoslemiM.Application of He's variational iteration method and Adomian's decomposition method to the fractional KdV-Burgers-Kuramoto equation20095811-122091209710.1016/j.camwa.2009.03.043MR2557332ZBL1189.65255SongL.ZhangH.Solving the fractional BBM-Burgers equation using the homotopy analysis method20094041616162210.1016/j.chaos.2007.09.042MR2531252ZBL1198.65205AbbasbandyS.ShirzadiA.Homotopy analysis method for multiple solutions of the fractional Sturm-Liouville problems201054452153210.1007/s11075-009-9351-7MR2670667ZBL1197.65096GepreelK. A.The homotopy perturbation method applied to the nonlinear fractional Kolmogorov-Petrovskii-Piskunov equations20112481428143410.1016/j.aml.2011.03.025MR2793646ZBL1219.35347JafariH.MomaniS.Solving fractional diffusion and wave equations by modified homotopy perturbation method20073705-638839610.1016/j.physleta.2007.05.118MR2419064ZBL1209.65111MitkowskiW.Approximation of fractional diffusion-wave equation2011526568ZhangX.LiuJ.WenJ.TangB.HeY.Analysis for one-dimensional time-fractional Tricomi-type equations by LDG methods201363114316410.1007/s11075-012-9617-3MR3040407SakarM. G.ErdoganF.YldrmA.Variational iteration method for the time-fractional Fornberg-Whitham equation2012639138213882-s2.0-8485975625310.1016/j.camwa.2012.01.031LiuC.-S.A new trial equation method and its applications20064533953972-s2.0-3364550157910.1088/0253-6102/45/3/003LiuC.-S.Trial equation method to nonlinear evolution equations with rank inhomogeneous: mathematical discussions and its applications20064522192232-s2.0-3364495873810.1088/0253-6102/45/2/005ChengY.-J.Classification of traveling wave solutions to the Vakhnenko equations201162103987399610.1016/j.camwa.2011.09.060MR2852116ZBL1236.35145GurefeY.SonmezogluA.MisirliE.Application of the trial equation method for solving some nonlinear evolution equations arising in mathematical physics2011776102310292-s2.0-8455520324310.1007/s12043-011-0201-5BulutH.KilicB.Exact solutions for some fractional nonlinear partial differential equations via Kudryashov method2013812431BulutH.KilicB.Finding exact solution by using a new auxiliary equation for fractional RLW burges equation201381110PandirY.GurefeY.KadakU.MisirliE.Classification of exact solutions for some nonlinear partial differential equations with generalized evolution201220121610.1155/2012/478531478531MR2965451ZBL1247.35124GurefeY.MisirliE.SonmezogluA.EkiciM.Extended trial equation method to generalized nonlinear partial differential equations2013219105253526010.1016/j.amc.2012.11.046MR3009484PandirY.GurefeY.MisirliE.Classification of exact solutions to the generalized Kadomtsev-Petviashvili equation201387112JafariH.SeifiS.Homotopy analysis method for solving linear and nonlinear fractional diffusion-wave equation20091452006201210.1016/j.cnsns.2008.05.008MR2474460ZBL1221.65278SahadevanR.BakkyarajT.Invariant analysis of time fractional generalized Burgers and Korteweg-de Vries equations2012393234134710.1016/j.jmaa.2012.04.006MR2921677ZBL1245.35142JumarieG.Modified Riemann-Liouville derivative and fractional Taylor series of nondifferentiable functions further results2006519-101367137610.1016/j.camwa.2006.02.001MR2237634ZBL1137.65001JumarieG.Fractional Hamilton-Jacobi equation for the optimal control of nonrandom fractional dynamics with fractional cost function2007231-221522810.1007/BF02831970MR2282462ZBL1111.49014JumarieG.Table of some basic fractional calculus formulae derived from a modified Riemann-Liouville derivative for non-differentiable functions200922337838510.1016/j.aml.2008.06.003MR2483503ZBL1171.26305SachdevP. L.Srinivasa RaoCh.N-wave solution of modified Burgers equation20001341610.1016/S0893-9659(99)00199-8MR1752130ZBL0957.35115BiswasA.TrikiH.HayatT.AldossaryO. M.1-soliton solution of the generalized Burgers equation with generalized evolution201121724102891029410.1016/j.amc.2011.05.031MR2806416ZBL1219.35221