Skip to content
BY 4.0 license Open Access Published by De Gruyter Open Access September 14, 2019

Lie symmetry analysis and conservation law for the equation arising from higher order Broer-Kaup equation

  • Hengtai Wang , Huiwen Chen , Zigen Ouyang EMAIL logo and Fubin Li
From the journal Open Mathematics


In this paper, Lie symmetry analysis is performed for the equation derived from $(2+1)$-dimensional higher order Broer-Kaup equation. Meanwhile, the optimal system and similarity reductions based on the Lie group method are obtained. Furthermore, the conservation law is studied via the Ibragimov’s method.

MSC 2010: 35L65; 37K05; 70S10

1 Introduction

Nonlinear partial differential equations (PDEs) arising in many physical fields like the condense matter physics, plasma physics, fluid mechanics and optics and so on. In order to investigate the exact solution of PDEs, a fruitful techniques have been developed, such as traveling wave transformations, inverse scattering method [1], Darboux and Bäcklund transformations [2], Lie symmetry analysis [3, 4, 5]. Lie symmetry analysis is a very useful method to find the new solutions of PDEs, which was distribution by Sophus Lie (1842-1899). In addition on the base of symmetries, the integrability of the nonlinear PDEs, such as group classification, optimal system and conservation laws, can be considered. Lie groups, as a type of trandformation groups, can tranfer one solution to another one of a given PDE. In other words, if we get one solution of a PDE, we can obtain the other ones via the symmetry of the PDE. Based on this, we will investigate the Lie symmetry analysis of the given PDE.

Noether’s theorem [6] establishes a connection between symmetries of differential equations and conservation laws. However, there are other methods to study the conservation laws, such as partial Noether’s approach, multiplier approach and Ibragimov’s method. As stated in [7], the former three methods are not applicable to the nonlinear PDEs that do not admit a Lagrangian. In order to overcome these difficulties, Ibragimov’s method was proposed [8]. Especially state, on the contribution of Lie symmetry method, significant researches have been done on the integrability of the nonlinear PDEs, group classification, optimal system, reduced solutions and conservation laws, such as [9, 10, 11, 12, 13, 14] and [15, 16, 17, 18, 19, 20, 21] published this year and last year.

The (2+1)-dimensional higher order Broer-Kaup equation was considered in [22] and [23], whose expression is as follows:

Ut+4(Uxx+U33UUx+3UW+3P)x=0,Vt+4(Vxx+VU2+UVx+3VW)x=0,WyVx=0,Py(UV)x=0. (1.1)

Li et al. and Mei et al. took the Bäcklund transformation of system (1.1) and obtained the relationship:


Such that (1.1) becomes a single differential equation:

Ut+4(Uxx+U3+3UUx)x=0. (1.2)

For (1.2), we consider its special case. That is, U = U(x, t) is regarded as (1 + 1)-dimensional and replaced by u, then (1.2) becomes

ut+4(uxx+u3+3uux)x=0. (1.3)

For convenience to cite later, we call (1.3) to be Li-Mei system, which is equivalent to


The exact traveling wave solutions have been investigated in [24]. However, to the best of our knowledge, the Lie symmetry, optional system and conservation law of Li-Mei equation have not been researched, which is the original intention of this work.

This paper is organized as follows. In section 2, we perform Lie symmetry analysis of Li-Mei system. In section 3, the optimal system and similarity reductions are studied. section 4 distributes to studying the conservation law in the method of Ibragimov’s and construction the conserved vectors.

2 Lie symmetries of Li-Mei equation (1.3)

Lie symmetries analysis will be performed of Eq. (1.3) in this section. Consider a one-parameter Lie group of transformations:

xx+εξ(x,t,u)+O(ε2),tt+ετ(x,t,u)+O(ε2),uu+εϕ(x,t,u)+O(ε2), (2.1)

With a small parameter ε ≪ 1. The vector field associated with the above transformation group can assumed as:

V=ξ(x,t,u)x+τ(x,t,u)t+ϕ(x,t,u)u (2.2)

Thus the third prolongation pr(3)V is:

pr(3)V=V+ϕxux+ϕtut+ϕxxuxx+ϕxxxuxxx, (2.3)

where only the terms involved in (1.3) appear in (2.3). In (2.3), ϕx, ϕt, ϕxx and ϕxxx are all undetermined functions, which are given by the following formulae.

ϕx=Dx(ϕξuxτut)+ξuxx+τuxt, (2.4)
ϕt=Dt(ϕξuxτut)+ξuxt+τutt, (2.5)
ϕxx=Dx2(ϕξuxτut)+ξuxxx+τuxxt, (2.6)
ϕxxx=Dx3(ϕξuxτut)+ξuxxxx+τuxxxt, (2.7)

where Dx, Dt are denoted the total derivatives with respect to x and t, respectively.

The determining equation of Eq. (1.3) arises from the following invariance condition:

pr(3)V(Δ)|Δ=0=0. (2.8)


Δ=ut+4uxxx+12u2ux+12ux2+12uuxx. (2.9)

By (2.8), we have the following symmetry condition:

ϕt+4ϕxxx+24ϕuux+12u2ϕx+24ϕxux+12ϕuxx+12uϕxx=0, (2.10)

which ξ(x, t, u), τ(x, t, u) and ϕ(x, t, u) must satisfy.

Substituting (2.4)-(2.7) into (2.10), replacing ut by −(4uxxx + 12u2 ux + 12ux2 + 12uuxx) whenever it appears, and comparing the coefficients of the various monomials in the first, second and third order partial derivatives, and solving the system, we obtain the expression of ξ(x, t, u), τ(x, t, u) and ϕ(x, t, u).

ξ(x,t,u)=c1x+c2,τ(x,t,u)=3c1x+c3,ϕ(x,t,u)=c1u, (2.11)

where c1, c2, c3 are arbitrary constants.

Hence the infinitesimal generators of Eq. (1.3) can be listed as follows

V1=x,V2=t,V3=xx+3ttuu. (2.12)

By solving the following ordinary differential equations with initial condition:

dxdε=ξ(x,t,u),x|ε=0=x,dtdε=τ(x,t,u),t|ε=0=t,dudε=ϕ(x,t,u),u|ε=0=u. (2.13)

We therefore obtain the group transformation which is generated by the infinitesimal generators V1, V2, V3, respectively:

G1:(x,t,u)(x+ε,t,u), (2.14)
G2:(x,t,u)(x,t+ε,u), (2.15)
G3:(x,t,u)(eεx,e3εt,eεu). (2.16)

Here G1, G2, G3 are all one-dimensional Lie groups generated by their own generators gi,ε, whose operation is manifested by (2.14), (2.15), (2.16), respectively.

It is trivial that V1, V2, V3 form a 3-dimensional Lie algebra L with the following Lie bracket:

[V1,V2]=0,[V1,V3]=V1,[V2,V3]=3V2. (2.17)

Remark 1

In (2.14)-(2.16), an arbitrary element in Gi(i = 1, 2, 3) can transfer one solution of Eq. (1.3) to another one, so do the products of the elements from G1, G2 and G3.

Remark 2

The Lie group G1 × G2 is a normal Lie subgroup of G1G2G3. The Lie algebra generated by V1 and V2 is an ideal of L.

Theorem 1

The vector fields V1, V2 and V3 supply a representation of the Lie algebra


where the Lie bracket is

[x1,x2]=0,[x1,x3]=x1,[x2,x3]=3x2. (2.18)

The definition of representations of Lie algebras see [25].


It is suffice if we take the representation space to be the set of all the analytic functions and the linear mapping ρ : xiVi for i = 1, 2, 3. □

Remark 3

The vector fields V1 and V2 have trivial prolongation. However, the prolongation of V3 can be computed:

pr(3)V3=xx+3ttuu4utut2uxux3uxxuxx4uxxxuxxx. (2.19)

It is easy to check pr(3)V3(Δ) = −4⋅Δ, which is called the symmetry invariance of differential equation (1.3).

We are now to take an example to illustrate the applications of Lie symmetry analysis. We take ut = uxx as an example rather than Eq. (1.3) since it is difficult to find the analytical solution. The vector fields of this equation is V1 = x, V2 = t, V3 = u, V4 = x/2x + tt + u∂u. It is not difficult to find a special solution u(x, t) = et(ex + ex). Under the operation of Lie group generated by V1V4, we can check that


are all the solutions of ut = uxx.

3 Optimal system of one-dimensional subalgebras

The more technical matters arose in order to classify the subalgebra of Lie algebra generated by Lie point symmetries, for instance [26] and [3]. A concise method to get the optimal system was presented by Ibragimov in 2010 [27]. In this section we shall construct an optimal system of one-dimensional subalgebra.

Theorem 2

The following operators provide two optimal systems of one-dimensional subalgebras of the Lie algebra spanned by V1, V2, V3 of Eq. (1.3):

I:V1,νV1+V2,V3, (3.1)


II:V2,V1+μV2,V3, (3.2)

where both ν and μ are arbitrary constants.


Suppose W and V are two vector field and


By solving this ODE we have

W(ε)=Ad(exp(εV))W0, (3.3)

by summing the Lie series[3]

Ad(exp(εV))W0=n=0εnn!(adV)n(W0)=W0ε[V,W0]+ε22![V,[V,W0]]. (3.4)

In view of (3.4), we obtain

Ad(exp(εVi))Vi=Vi,i=1,2,3;Ad(exp(εV1))V2=V2,Ad(exp(εV1))V3=V3εV1;Ad(exp(εV2))V1=V1,Ad(exp(εV2))V3=V33εV2;Ad(exp(εV3))V1=eεV1,Ad(exp(εV3))V2=e3εV2. (3.5)

For an arbitrary nonzero vector


our task is to simplify as many of the coefficients ai as possible through the applications of adjoint maps to V.

  1. a3 ≠ 0. Scaling V if necessary, we can assume that a3 = 1. By making use of (3.5) and acting on such a V by Ad(exp(εa23V2)), we can make the coefficient of V2 vanish:


    Next we act on V′ by Ad(exp(ε a1V1)), to cancel the coefficient of V1. Hence V is equivalent to V3 under the adjoint representation.

  2. a3 = 0.

  3. a2 ≠ 0, a1 ≠ 0. Without losing generality, we can assume that a2 = 1. One can easily figure out that the adjoint representation induced by any combinations of V1, V2, V3 shall make a1V1 + V2 invariant. In other words, any one-dimensional subalgebra generated by V is equivalent to the subalgebra generated by a1V1 + V2.

  4. a2 = 0, a1 ≠ 0. Similarly to the discussion of Subcase 1, we can conclude that V is equivalent to V1 under the adjoint representation.

The other optimal system can be obtained similarly. □

4 Similarity reductions and exact solutions for Eq. (1.3)

In the preceding section, we got the optimal system of Eq. (1.3). We are now in the position to deal with the symmetry reduction and exact solutions via constructing similarity variables.

  1. For the generator V1, we assume ζ = t, u = f(ζ) and the we obtain the trivial solution f = c, where c is an arbitrary constant.

  2. For the linear combination ν V1 + V2, we have

    u=f(ζ), (4.1)

    where ζ = xν t, which is a traveling wave transformation. By substituting (4.1) into Eq. (1.3), we reduce this equation to the following ODE

    4f+12ff+12f2+12f2fνf=0, (4.2)

    where f=dfdζ,ν0.

    The traveling wave solutions were obtained in [24].

  3. For the generator V3, we have

    u=t13f(ζ), (4.3)

    where ζ = xt−1/3. Substituting (4.3) into Eq. (1.3), we reduce it to the following ODE

    4f+12ff+12f2+12f2f13ζf13f=0, (4.4)

    where f=dfdζ.

    For optimal system II, we only discuss the similarity reductions of V2 and V1 + μ V2.

  4. For the generator V2, we have

    u=f(ζ), (4.5)

    where ζ = x. By substituting (4.5) into Eq. (1.3), we reduce this equation to the following ODE

    4f+12ff+12f2+12f2f=0, (4.6)

    where f=dfdζ.

  5. For the linear combination V1 + μ V2, we have

    u=f(ζ), (4.7)

    where ζ = μ xt, which follows that this ODE

    4μ3f+12μ2ff+12μ2f2+12μf2ff=0, (4.8)

    where f=dfdζ and μ ≠ 0.

Figure 1 
The graph of f(ζ) given by Eq. (4.2) as ν takes −5, 0, 5.
Figure 1

The graph of f(ζ) given by Eq. (4.2) as ν takes −5, 0, 5.

Figure 2 
The graph of f(ζ) given by Eq. (4.4).
Figure 2

The graph of f(ζ) given by Eq. (4.4).

Figure 3 
The graph of f(ζ) given by Eq. (4.6).
Figure 3

The graph of f(ζ) given by Eq. (4.6).

Figure 4 
The graph of f(ζ) given by Eq. (4.8) for μ = −10, −5, 5, 10.
Figure 4

The graph of f(ζ) given by Eq. (4.8) for μ = −10, −5, 5, 10.

Figure 5 
The graph of u(x, t) given by Eq. (4.3) and Eq. (4.4).
Figure 5

The graph of u(x, t) given by Eq. (4.3) and Eq. (4.4).

In the above, we sketch the graphs of f(ζ) in Eqs. (4.2), (4.4), (4.6), (4.8) and 3D-plot of u(x, t) in Eqs. (4.3), (4.4) under the initial conditions f(0) = 32 , f(1) = 1, f′(0) = 0.

5 Nonlinear self-adjointness and conservation law

First of all we show that Li-Mei equation is nonlinearly self-adjoint.

For a given PDEs

Rβ(x,u,u(1),,u(k))=0, (5.1)

define the Euler-Lagrange operator

δδuαuα+j=1(1)jDi1Dijui1ijα,α=1,2,,m, (5.2)

and the formal Lagrangian

L=β=1mvβRβ(x,u,u(1),,u(k)). (5.3)

The adjoint equations

(Rα)(x,u,u(1),,u(k))=δLδuα=0,α=1,2,,m,v=v(x). (5.4)

Definition 1

The system (5.1) is said to be nonlinearly self-adjoint if the adjoint system (5.4) is satisfied for all solutions u of system (5.1) upon a substitution v = φ(x, u) such that φ(x, u) ≠ 0, which is equivalent to the following identity holding for the undetermined functions λαβ

(Rα)(x,u,u(1),v(1),,u(k),v(k))|v=φ(x,uα)=β=1mλαβRβ. (5.5)

In this paper α = 1 and Rα(x, u, u(1), ⋯, u(k)) = Δ(x, t, ut, ux, uxx, uxxx) = Eq. (2.9). The formal Lagrangian is

L=v(ut+4uxxx+12u2ux+12ux2+12uuxx). (5.6)

Substituting it into (5.2) = 0, we have the adjoint equation to Eq. (1.3)

4vxxx12uvxx+12u2vx+vt=0. (5.7)

By means of

(4vxxx12uvxx+12u2vx+vt)|v=φ(x,t,u)=λΔ, (5.8)

it leads us to

12u2φx12u2φuuxφtφuut+12uφxx+24uφuxux+12uφuuxx12φuuux24φxxx12φuxxux12φuuxux212φuxuxx4φuuuux312φuuuxuxx4φuuxx=λ(ut+4uxxx+12u2ux+12ux2+12uuxx) (5.9)

Firstly we obtain λ = −φu by comparing the terms with the third-order derivative of u. And then

φuux+φuuφu=0φuu=0. (5.10)

We therefore get φu = 0 and 4φxxx − 12xx + 12u2φx + φt = 0. In view of φu = 0, one has φx = φt = 0, so φ = c ≠ 0, which proves that Eq. (1.3) is self-adjoint.

We are now in the position to construct the conservation law of Eq. (1.3).

Theorem 3

(Ibragimov’s method). Let the system of differential Eq. (5.1) be nonlinearly self-adjoint. Then every Lie point, Lie-Bäcklund, nonlocal symmetry

X=ξi(x,u,u(1),)xi+ηα(x,u,u(1),)uα (5.11)

admitted by the system of Eq. (5.1) gives rise to a conservation law, where the components 𝓒i of the conserved vector 𝓒 = (𝓒1, ⋯, 𝓒n) are determined by

Ci=WαLuiαj=1nDjLuijα+j,k=1nDjDkLuijkα+j=1nDj(Wα)Luijαk=1nDkLuijkα+j,k=1nDjDk(Wα)Luijkα, (5.12)

with Wα=ηαj=1nξjujα.

For the generator V=ξx+τt+ϕu, we have W = ϕξ uxτ ut, we therefore obtain the following components of conserved vector

Cx=WLuxDxLuxx+Dx2Luxxx+Dx(W)LuxxDxLuxxx+Dx2(W)Luxxx, (5.13)
Ct=WLut. (5.14)

Taking the formal Lagrangian 𝓛 given by (5.6) into (5.13) and (5.14), we can simplify the expressions of 𝓒x and 𝓒t as follows

Cx=W12u2v+24vuxDx(12uv)+Dx2(4v)+Dx(W)12uvDx(4v)+Dx2(W)(4v), (5.15)
Ct=Wv. (5.16)

For the generator V1 = x , it has the Lie characteristic function W = −ux. By using of the formulae (5.15) and (5.16), it can give rise to the following components of the conserved vector


For the generator V2 = t , we have W = −ut, the formulae (5.15) and (5.16) yield the following components of the conserved vector


For the generator V3=xx+3ttuu, we have W = − uxux − 3tut, the formulae (5.15) and (5.16) imply the following components of the conserved vector


These vectors involve an arbitrary solution v of the adjoint equation (5.7) and hence provide an infinite number of conservation laws.

6 Conclusions

In this paper, we have obtained the symmetries and the corresponding Lie algebras of Li-Mei system by using Lie symmetry analysis method. Meanwhile, the optimal system and its similarity reductions are investigated. Furthermore, we proved that it is nonlinearly self-adjoint. Finally, the conserved vectors were constructed via the Ibragimov’s method.

The vector fields generate the equation under consideration supply a representation of a Lie algebra. However, for a given finitely dimensional Lie algebra, such as nine types of simply Lie algebras, how to get its representation via vector fields? If we have already obtained the vector fields, can we get the differential equation which generates the vector field? If the differential equation is obtained, is it unique? All of them are the aims that we will study in the near future.

  1. Authors’ contributions Hengtai Wang denotes to studying of the whole system and writing the main body of the article. Huiwen Chen contributes the graphs of the paper. Zigen Ouyang checks all the errors of this paper. Fubin Li denotes some calculations of this article.


The authors are grateful to the editor for his/her valuable comments and suggestions. This work is supported by the National Natural Science Foundation of China (Grant No. 11801264, 11601222) and Hunan Provincial Natural Science Foundation of China (Grant Nos. 2019JJ40240, 2019JJ50487).


[1] Gardner C.S., Greene J.M., Kruskal M.D., Miura R.M., Method for solving the Korteweg-de Vries equation, Phys. Rev. Lett., 1967, 19, 1095-1097.10.1103/PhysRevLett.19.1095Search in Google Scholar

[2] Li Y.S., Soliton and integrable systems, Advanced Series in Nonlinear Science, Shanghai Scientific and Technological Education Publishing House, Shang Hai, 1999 (in Chinese).Search in Google Scholar

[3] Olver P.J., Applications of Lie groups to differential equations, Grauate Texts in Mathematics, Springer, New York, 1993, 107.10.1007/978-1-4612-4350-2Search in Google Scholar

[4] Bluman G.W., Kumei S., Symmetries and Differential Equations, Springer-Verlag, World Publishing Corp., 1989.10.1007/978-1-4757-4307-4Search in Google Scholar

[5] Cantwell B.J., Introduction to Symmetry Analysis, Cambridge University Press, 2002.Search in Google Scholar

[6] Noether E., Invariante Variationsprobleme, Königliche Gesellschaft der Wissenschaften zu Göttingen, Nachrichten, Mathematisch-Physikalische Klasse Heft 1918, 2, 235-257, English transl.: Transport Theory Statist. Phys. 1971, 1, 186-207.Search in Google Scholar

[7] Zhao Z.L., Han B., Lie symmetry analysis of the Heisenberg equation, Commun. Nonlinear Sci. Numer. Simulat., 2017, 45, 220-234.10.1016/j.cnsns.2016.10.008Search in Google Scholar

[8] Ibragimov N.H., A new conservation theorem, J. Math. Anal. Appl., 2007, 333, 311-328.10.1016/j.jmaa.2006.10.078Search in Google Scholar

[9] Liu H.Z., Li J.B., Lie symmetry analysis and exact solutions for the short pulse equation, Nonlinear. Anal., 2009, 71, 2126-2133.10.1016/ in Google Scholar

[10] Liu H.Z., Li J.B., Liu L., Lie group classifications and exact solutions for two variable-coefficient equations, Appl. Math. Comput., 2009, 215, 2927-2935.10.1016/j.amc.2009.09.039Search in Google Scholar

[11] Liu H.Z., Li J.B., Lie symmetries, conservation laws and exact solutions for two Rod equations, Acta. Appl. Math., 2010, 110, 573-587.10.1007/s10440-009-9462-0Search in Google Scholar

[12] Liu H.Z., Li J.B., Liu L., Lie symmetry analysis, optimal systems and exact solutions to the fifth-order KdV types of equations, J. Math. Anal. Appl., 2010, 368, 551-558.10.1016/j.jmaa.2010.03.026Search in Google Scholar

[13] Liu H.Z., Li J.B., Liu L., Group classifications, symmetry reductions and exact solutions to the nonlinear elastic Rod equations, Adv. Appl. Clifford Algebras, 2012, 22, 107-122.10.1007/s00006-011-0290-8Search in Google Scholar

[14] Nadjafikhah M., Ahangari F., Symmetry analysis and conservation laws for the Hunter-Saxton equation, Commun. Theor. Phys., 2013, 59, 335-348.10.1088/0253-6102/59/3/16Search in Google Scholar

[15] Paliathanasis A., Tsamparlis M., Lie symmetries for systems of evolution equations, J. Geom. Phys., 2018, 124, 165-169.10.1016/j.geomphys.2017.10.014Search in Google Scholar

[16] Rashidi S., Hejazi S.R., Lie symmetry approach for the Vlasov-Maxwell system of equations, J. Geom. Phys., 2018, 132, 1-12.10.1016/j.geomphys.2018.04.014Search in Google Scholar

[17] Bansal A., Biswas A., Zhou Q., Babatin M. M., Lie symmetry analysis for cubic-quartic nonlinear Schrodinger’s equation, Optik, 2018, 169, 12-15.10.1016/j.ijleo.2018.05.030Search in Google Scholar

[18] Dorjgotov K., Ochiai H., Zunderiya U., Lie symmetry analysis of a class of time fractional nonlinear evolution systems, Appl. Math. Comput., 2018, 329, 105-117.10.1016/j.amc.2018.01.056Search in Google Scholar

[19] Kumar M., Tiwari A.K., Soliton solutions of BLMP equation by Lie symmetry approach, Comput. Math. Appl., 2018, 75, 1434-1442.10.1016/j.camwa.2017.11.018Search in Google Scholar

[20] Zhang Y., Zhai X.H., Perturbation to Lie symmetry and adiabatic invariants for Birkhoffian systems on time scales, Commun. Nonlinear Sci. Numer. Simulat., 2019, 75, 251-261.10.1016/j.cnsns.2019.04.005Search in Google Scholar

[21] Kumar M., Tanwar D.V., On Lie symmetries and invariant solutions of (2 + 1)-dimensional Gardner equation, Commun. Nonlinear Sci. Numer. Simulat., 2019, 69, 45-57.10.1016/j.cnsns.2018.09.009Search in Google Scholar

[22] Li D., Gao F., Zhang H., Solving the (2 + 1)-dimensional highter order Broer-Kaup system via a transformation and tanh-function method, Chaos Solit. Fract., 2004, 20, 1021-1025.10.1016/j.chaos.2003.09.006Search in Google Scholar

[23] Mei J., Li D., Zhang H., New soliton-like and periodic solution of (2 + 1)-dimesional highter order Broer-Kaup system, Chaos Solit. Fract., 2004, 22, 669-674.10.1016/j.chaos.2004.02.023Search in Google Scholar

[24] Li J.B., On the exact traveling wave solutions of (2 + 1)-dimensional higher order Broer-Kaup equation, Int. J. Bifurcat. Chaos, 2014, 24, 1450007.10.1142/S0218127414500072Search in Google Scholar

[25] Humphreys J.E., Introduction to Lie algebras and representation theory, Springer-Verlag, New York, 1972.10.1007/978-1-4612-6398-2Search in Google Scholar

[26] Ovsiannikov L.V., Group analysis of differential equations, Academic Press, 1982.10.1016/B978-0-12-531680-4.50012-5Search in Google Scholar

[27] Grigoriev Y.N., Ibragimov N.H., Kovalev V.F., Meleshko S.V., Symmmetry of integro-differential equations: with applications in mechanics and plasma physica, Springer, 2010.10.1007/978-90-481-3797-8Search in Google Scholar

Received: 2018-09-28
Accepted: 2019-05-21
Published Online: 2019-09-14

© 2019 Wang et al., published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 Public License.

Downloaded on 26.3.2023 from
Scroll Up Arrow