Documentos de Académico
Documentos de Profesional
Documentos de Cultura
Abstract. In this letter, the Lie point symmetries of a class of Gordon-type wave equations that
arise in the Milne space-time are presented and analysed. Using the Lie point symmetries, it is
showed how to reduce Gordon-type wave equations using the method of invariants, and to obtain
exact solutions corresponding to some boundary values. The Noether point symmetries and conser-
vation laws are obtained for the Klein–Gordon equation in one case. Finally, the existence of higher-
order variational symmetries of a projection of the Klein–Gordon equation is investigated using the
multiplier approach.
1. Introduction
A vast amount of work has been published in the literature studying differential equations
(DEs) in terms of the Lie point symmetries admitted by them [1,2]. These symmetries
play an important role in finding exact analytic solutions of the nonlinear DEs. Other than
Lie point symmetries, Noether symmetries are also widely studied and are associated, in
particular, with those DEs which possess Lagrangians. These symmetries represent phys-
ical features of DEs via the conservation laws they admit. The interesting link between
symmetries and conservation laws in mathematical physics is provided in the classic work
of Noether [3] showing that for every infinitesimal transformation admitted by the action
integral of a system, there exists a conservation law. The Noether symmetries, which are
symmetries of the Euler–Lagrange systems, have interesting applications in the study of
properties of particles moving under the influence of gravitational fields.
where u (1) , u (2) , . . . , u (r ) denote the collections of all first-, second-, . . ., r th-order par-
tial derivatives, that is, u iα = Di (u α ), u iαj = D j Di (u α ), . . . respectively, with the total
differentiation operator with respect to x i given by
∂ ∂ ∂
Di = + u iα α + u iαj α + · · · , i = 1, . . . , n, (2)
∂x i ∂u ∂u j
Di i = 0 (3)
along the solutions of (1). It can be shown that every admitted conservation law arises
from multipliers Q μ (x, u, u (1) , . . .) such that
Q μ G μ = Di i (4)
holds identically (that is, off the solution space) for some current . The conserved
vector may then be obtained by the homotopy operator (see [1,9,10]). Other works on
symmetries and conservation laws can be found in [11–13].
DEFINITION 1
Note. In most literature, a variational problem consists of finding the extrema (maxima
or minima) of a functional
L[u] = L(x, u (n) )dx
DEFINITION 2
ζiα = Di (W α ) + ξ j u iαj ,
ζiα1 ...is = Di1 . . . Dis (W α ) + ξ j u αji1 ...is , s > 1. (7)
DEFINITION 3
DEFINITION 4
where the Euler–Lagrange operators with respect to derivatives of u α are obtained from
(5) by replacing u α by the corresponding derivatives, e.g.,
δ ∂ ∂
α = α + (−1)s D j1 · · · D js α ,
δu i ∂u i s≥1
∂u i j1 ··· js
i = 1, . . . , n, α = 1, . . . , m. (11)
i = f i − N i (L) , i = 1, . . . , n, (12)
which represents an empty Universe and is of interest in relativity for being a special
case of a well-known Friedmann–Lemaître–Robertson–Walker metric [6,8]. The Klein–
Gordon equation [8] on (13) is obtained by
1 ∂ ij ∂
u = √ |−g|g u = k(u), (14)
|−g| ∂ x i ∂xi
and takes the form
In order to find the Lie point symmetries of the above Gordon-type equation we restrict
k(u) to some special cases. These cases are assumed by keeping in mind the fact that we
allow the inhomogeneous term, k(u), to be taken as sin(u) and some powers of u. The
criterion that yields the Lie point is given by the invariance condition [2]
where X is the prolonged symmetry generator in the jet space. Thus, the invariance
of differential equations (15) leads to the Lie point symmetries possessed by (15). The
procedure for finding Lie point symmetries is well known [2] and therefore will be given
without derivations. It turns out, from the symmetry study, that some special polynomial
cases of k(u) arise. Also, in line with the literature, we consider the sine-Gordon equation.
Thus, we study the four cases for k(u) in (15) given by
(i) k(u) = sin(u) (sine-Gordon)
(ii) k(u) = u (Klein–Gordon)
(iii) k(u) = u 3
(iv) k(u) = u n , n = 0, 1, 3.
Case i. Following the symmetry criterion, we find that eq. (16) in this case admits ten Lie
point symmetries given by
ex
X1 = ∂ x − e x ∂t ,
t
X 2 = ∂y ,
e−x ex y
X3 = ∂ y − ex y∂t + ∂x ,
t t
X 4 = ∂z ,
e−x ex z
X5 = ∂z − ex z∂t + ∂x ,
t t
X 6 = y∂z − z∂ y ,
X 7 = −∂x + y∂ y + z∂z ,
2e−x y 2e−x z
X8 = ∂y + ∂z + e−x (−1 − e2x (y 2 + z 2 ))∂t
t t
e−x (−1 + e2x (y 2 + z 2 ))
+ ∂x ,
t
X 9 = 2y∂x − 2yz∂z + (e−2x − y 2 + z 2 )∂ y ,
X 10 = −2yz∂ y + 2z∂x + (e−2x + y 2 − z 2 )∂z .
Case ii. When k(u) = u, we have a Klein–Gordon equation. Equation (16) in this case
admits 13 Lie point symmetries given by
X 1 = u∂u ,
X 2 = F1 (x, y, z, t)∂u ,
ex
X3 = ∂ x − e x ∂t ,
t
X 4 = ∂y ,
e−x ex y
X5 = ∂ y − ex y∂t + ∂x ,
t t
X 6 = y∂z − z∂ y ,
X 7 = −∂x + y∂ y + z∂z ,
X 8 = ∂z ,
e−x ex z
X9 = ∂z − ex z∂t + ∂x ,
t t
X 10 = −2∂x + 2y∂ y + 2z∂z + u∂u ,
2e−x y 2e−x z
X 11 = ∂y + ∂z + e−x (−1 − e2x (y 2 + z 2 ))∂t
t t
e−x (−1 + e2x (y 2 + z 2 ))
+ ∂x ,
t
X 12 = 2y∂x − 2yz∂z + (e−2x − y 2 + z 2 )∂ y ,
X 13 = −2yz∂ y + 2z∂x + (e−2x + y 2 − z 2 )∂z ,
where
Case iii. When k(u) = u 3 , the (Gordon-type) eq. (16) yields a set of 15 Lie point
symmetries:
X 1 = t∂t − u∂u ,
ex (−1 + t 2 )
X 2 = −2ex tu∂u + ex (1 + t 2 )∂t + ∂x ,
t
ex (1 + t 2 )
X 3 = −2ex tu∂u + ex (−1 + t 2 )∂t + ∂x ,
t
X 4 = ∂y ,
e−x (−1 + t 2 ) ex (−1 + t 2 )y
X 5 = −2ex t yu∂u + ∂ y + ex (1 + t 2 )y∂t + ∂x ,
t t
X 6 = y∂z − z∂ y ,
X 7 = −∂x + y∂ y + z∂z ,
e−x (1 + t 2 ) ex (1 + t 2 )y
X 8 = −2ex t yu∂u + ∂ y + ex (−1 + t 2 )y∂t + ∂x ,
t t
X 9 = ∂z ,
e−x (1 + t 2 ) ex (1 + t 2 )z
X 10 = 2ex t zu∂u − ∂z − ex (−1 + t 2 )z∂t − ∂x ,
t t
Case iv. In this case k(u) = u n , n = 0, 1, 3, the general polynomial Gordon-type equation
(16) admits 11 Lie point symmetries given by
2u t − nt
X1 = ∂u + ∂t ,
−3 + n −3 + n
ex
X2 = ∂ x − e x ∂t ,
t
X 3 = ∂y ,
e−x ex y
X4 = ∂ y − ex y∂t + ∂x ,
t t
X 5 = ∂z ,
X 6 = y∂z − z∂ y ,
X 7 = −∂x + y∂ y + z∂z ,
e−x ex z
X8 = ∂z − ex z∂t + ∂x ,
t t
2e−x y 2e−x z
X9 = ∂y + ∂z + e−x (−1 − e2x (y 2 + z 2 ))∂t
t t
e−x (−1 + e2x (y 2 + z 2 ))
+ ∂x ,
t
X 10 = −2y∂x + 2yz∂z + (−e−2x + y 2 − z 2 )∂ y ,
X 11 = 2yz∂ y − 2z∂x + (−e−2x − y 2 + z 2 )∂z .
In this section we briefly show how the order of (1+3) Klein–Gordon equation (15) can
be reduced using its symmetries. In the first reduction, the equation with four independent
variables is reduced to a partial DE that has two independent variables. The reduced equa-
tion may then be analysed further using another Lie symmetry reduction or an appropriate
alternative method. In the second reduction (§2.1.1), we obtain exact solutions.
Since [X 6 , X 7 ] = 0, where X 6 and X 7 appear as Lie symmetries in all the above cases,
we may begin reducing with either X 6 or X 7 . Suppose we reduce (15) by X 6 = y∂z − z∂ y .
The characteristic equations are
dx dt dy dz du
= = = = .
0 0 −z y 0
1 2 3 2
u x x − u tt + 2 e−2x (αu αα + u α ) − u t + 2 u x − k(u) = 0 (17)
t2 t t t
with u = u(x, t, α).
If we then reduce eq. (17) by X 7 = −∂x + y∂ y + z∂z , we obtain the scaling
transformation X̄ = −∂x + 2α∂α . We now have the characteristic equations,
dt dx dα du
= = = .
0 −1 2α 0
By integrating, we obtain β = ln α + 2x and eq. (17) reduces to
2 4 3
u ββ (2 + e−β ) − u tt + 2 u β − u t − k(u) = 0 (18)
t2 t t
with u = u(t, β).
Equation (18) may be further analysed or reduced using the underlying symmetries.
It turns out that the Lie point symmetries are cumbersome and involve special functions
such as Bessel functions for k = u.
For k = u 3 , eq. (18) admits one symmetry
t∂t − u∂u .
3 1
X∗ = t∂t − u∂u ,
4 2
X ∗∗ = 4t 2 (1 + 2eβ )∂β + t 3 (1 + 4eβ )∂t − 2t 2 u(1 + 4eβ )∂u .
γ 4 Fγ γ + γ Fγ + 4F − 8γ 2 F 4 = 0, (19)
where
te−β/4
γ = and F = ueβ/2 (1 + 2eβ )1/2 .
(1 + 2eβ )1/4
It turns out that eq. (19) admits the Lie point symmetry
3
G = − β∂β + F∂ F
2
which leads us to the first-order ODE
2q + 24q 2/3 p 4/3 − 12q 2/3 p 1/3
qp = , (20)
2 p + 3q 1/3 p 2/3
where p = γ 2 F 3 and q = γ 5 F 3 .
There are no symmetries for k = u n , n = 0, 1, 3, 4 and k = sin u in eq. (18).
Also, one may consider reduction by studying the underlying conservation laws. This
would require methods other than the variational one, i.e., Noether’s theorem, since
eq. (18) is not variational.
For the Klein–Gordon case k(u) = u in (18), it can be shown, for e.g., that a conserved
vector of (18) is (β , t ), where
such that Dβ β + Dt t = 0 along the solutions of (18) and where BesselJ is the Bessel
function of the first kind ((β is the conserved flow and t , the conserved density).
For k(u) = u 3 in (18), the components of the conserved vector are:
β = (1 + 2eβ )t 2 [u t u β − uu βt ],
1
t = − t 2 [eβ t 2 u 4 + 2eβ t 2 u t 2 + 4u(eβ tu t − 2eβ u β − (1 + 2eβ )u ββ )].
4
Similarly, for k(u) = u 4 in (18), the components of the conserved vector are
1
β = − (1 + 2eβ )t 3 (40eβ u 2 + 4eβ t 2 u 5 − 5u β ((1 + 4eβ )tu t
5
+ 4(1 + 2eβ )u β ) + 5tu(10eβ u t + 2eβ tu tt + (1 + 4eβ )u βt )),
1 1
t = −t 4 (−2eβ (1 + 4eβ )u 2 + eβ (1 + 4eβ )t 2 u 5 + eβ tu t ((1 + 4eβ )tu t
5 2
+ 4(1 + 2eβ )u β ) − u(2eβ (3 + 8eβ )u β
+ (1 + 2eβ )(2eβ tu βt + (1 + 4eβ )u ββ ))).
The cases k(u) = sin u and k(u) = u n , n = 1, 3, 4 in eq. (18) do not yield any
conserved vectors.
2.1.1 Exact solutions/boundary conditions. We reduce eq. (15) by first using the
symmetry X = ∂ y to obtain
1 1 3 2
u x x − u tt + 2 e−2x u zz − u t + 2 u x − k(u) = 0,
t2 t t t
u = u(x, z, t), u(x, 0, t) = 0, u(x, 1, t) = 1. (21)
Reduce (21) further using the symmetry X = (e /t)∂x − e ∂t . The characteristic
x x
equations are
tdx dt dz du
= = = .
e x −e x 0 0
Integrating yields t˜ = (1/t)e−x and eq. (21) is reduced to
1
u zz − k(u) = 0, (22)
t˜2
where u(z, t˜) and boundary conditions transform to
u(0, t˜) = 0, u(1, t˜) = 1.
If k(u) = u, as in the Klein–Gordon case, then the general solution to eq. (22) is
u(z, t˜) = et˜z C1 (t˜) + e−t˜z C2 (t˜). (23)
We plot the function t˜ over different ranges: {x, −50, 50}, {t, 1, 10} (figure 1) and
{x, −10, 10}, {t, 1, 10} (figure 2). When t → ±∞, t˜ → 0, and when t → 0, t˜ becomes
large.
1 −x
Figure 1. t˜ = e .
t
1 −x
Figure 2. t˜ = e .
t
Keeping this in mind, we plot u(z, t˜) from (23) in figure 3: {t˜, 0, 10}, {z, −5, 5} and
figure 4: {t˜, 0, 10}, {z, −5, 5}, choosing particular C1 s and C2 s.
If we impose the boundary conditions on eq. (23), the solution of u(z, t˜) may be
expressed in terms of the hyperbolic sine function, namely,
1 1
u(z, t˜) = et˜z − e−t˜z .
˜
2 sinh(t ) 2 sinh(t˜)
The graph of this solution, for the range {t˜, 0, 10}, {z, −5, 5}, is given in figure 5.
Remark. For the sine-Gordon case, with k(u) = sin(u) in eq. (22), we obtain
where Jacobi amplitude [v, m] refers to the amplitude am(v m) for Jacobi elliptic
functions.
Consider the wave eq. (15) with k(u) = u (Klein–Gordon), which has the Lagrangian,
1 3 2x 2 1 2x 2 1 2 1 2 1 3 2x 2
L= t e u + te u x + tu y + tu z − t e u t . (24)
2 2 2 2 2
We assume that
is a Noether point operator that satisfies (9) with gauge vector f i (i = 1, 2, 3, 4), depen-
dent on (t, x, y, z, u). Then the Noether symmetry criterion (9) for the Lagrangian given
by eq. (24) takes the form,
X L + L[Dt τ + Dx ξ + D y η + Dz γ ] = Dt f 1 + Dx f 2 + D y f 3 + Dz f 4 .
u 3t : τu = 0,
u 3x : ξu = 0,
u 3y : ηu = 0,
u 3z : γu = 0,
3 1 1 1
u 2t : −e2x t 3 ξ − e2x t 2 τ − e2x t 3 γz − e2x t 3 η y − e2x t 3 ξx
2 2 2 2
1
+ e2x t 3 τt − e2x t 3 φu = 0,
2
1 1 1 1 1
u 2x : e2x tξ + e2x τ + e2x tγz + e2x tη y − e2x tξx + e2x tτt
2 2 2 2 2
+ e2x tφu = 0,
1 1 1 1 1
u 2y : τ + tγz − tη y + tξx + tτt + tφu = 0,
2 2 2 2 2
1 1 1 1 1
u 2z : τ − tγz + tη y + tξx + tτt + tφu = 0,
2 2 2 2 2
u t u z : e2x t 3 γt − tτz = 0,
u x u z : −e2x tγx − tξz = 0,
u y u z : −tγ y − tηz = 0,
u t u y : e2x t 3 ηt − tτ y = 0,
u y u x : −e2x tηx − tξ y = 0,
Consider the Klein–Gordon wave equation in Milne space-time with dependent variable
u as a function of x, t and y only, i.e., we have removed the spatial variable z from the
original wave equation (15) – the calculations are extremely cumbersome producing no
final outcomes. We consider the multiplier method for eq. (15), by choosing k(u) = u.
That is, since the Euler–Lagrange operator annihilates total divergences, we get
δ 1 1 3 2
Q 2 u x x − u tt + 2 e−2x u yy − u t + 2 u x − u = 0, (26)
δu t t t t
1 1 3 2
Q 2
u x x − u tt + 2 e−2x u yy − u t + 2 u x − u
t t t t
= D t t + D x x + D y y ,
where (x , y , t ) is the conserved flow (t being the conserved density). We obtain
the set of multipliers Qi , namely,
1 1 1
Q1 = t e 3 2x
− u x x x − yu y + yu x x y + y 2 u yy
3 6 2
1 1 1
+ y 3 u yyy − y 2 u x yy − u x x + u ,
3 2 12
Q2 = t 3 e2x (−2yu x yy + u x x y + yu yy + y 2 u yyy ),
1 3 −2x
Q3 = t e u yyy + t 3 u x x y
2
1
+ t e 2y 3 u x yy − y 3 u yy − y(u) − 3y 2 u x x y
3 2x
2
1 2 1 4
+ y u y + 3yu x x + 2yu x x x − y u yyy ,
2 2
1 3 2x
Q4 = t e (2u x yy − u yy − 2yu yyy ),
2
Q5 = t 3 e2x u yyy ,
1
Q7 = t e u x + u − yu y ,
3 2x
2
Q8 = t 3 e2x u y .
These multipliers yield a set of eight conserved flows, for example, the corresponding
components of the conserved vector for Q4 are
1
4x = t (2e2x t 2 u y 2 + 2u yy 2 − 6e2x tu yy u t − 2e2x t 2 u yy u tt
12
+ e2x u yy u x − 6e2x yu yyy u x + 6e2x u x u x yy + 2e2x u yy u x x
− 2u y (u yyy + e2x (−3tu t y − t 2 u tt y + 2u x y + u x x y ))
+ u(−4e2x t 2 u yy + 2u yyyy + e2x (−6tu t yy − 2t 2 u tt yy
+ 7u x yy + 6yu x yyy − 4u x x yy ))),
1 2x 3
t4 = e t (u yy u t + 2yu yyy u t − uu t yy − 2yuu t yyy − 2u t u x yy + 2uu xt yy ),
4
y 1
4 = t (−6e2x t 2 yu y 2 − 2(3yu yy 2 + u yyy u x
12
+ u yy (−9e2x t yu t − 3e2x t 2 yu tt + 6e2x yu x − 2u x y + 3e2x yu x x )
+ e2x (−3tu t y u x − t 2 u tt y u x + 6tu t u x y + 2t 2 u tt u x y
− 2u x u x y − 2u x y u x x + u x u x x y ))
+ u y (3e2x tu t − 18e2x t yu t y + e2x t 2 u tt
− 6e2x t 2 yu tt y − 2e2x u x + 4e2x t 2 u x + 12e2x yu x y + 4u x yy
+ 6e2x tu xt + 2e2x t 2 u xtt − 5e2x u x x + 6e2x yu x x y − 2e2x u x x x )
+ u(2e2x t 2 u y + 12e2x t 2 yu yy + 3e2x tu t y
+ 18e2x t yu t yy + e2x t 2 u tt y + 6e2x t 2 yu tt yy − 2e2x u x y
− 8e2x t 2 u x y − 12e2x yu x yy
− 2u x yyy − 12e2x tu xt y − 4e2x t 2 u xtt y + 7e2x u x x y
− 6e2x yu x x yy + 4e2x u x x x y )).
5. Concluding remarks
This paper investigates a class of wave and Gordon-type equations in Milne space-time.
In particular, we conducted a Lie and Noether symmetry analysis of a Klein–Gordon
equation on this manifold. We have given some symmetry reductions to show how the
(1+3)-dimensional wave equation can be reduced to an ordinary differential equation
using the method of invariants, and obtained some exact solutions. A conserved density
of the Klein–Gordon equation is constructed. Finally, some higher-order symmetries for
the projected equation and associated conservation laws are presented. It is hoped that an
analysis of the nonlinear Klein–Gordon equation in a genuinely curved space-time will
provide interesting insight from the point of view of conserved quantities.
References
[1] S Anco and G Bluman, Eur. J. Appl. Math. 13, 545 (2002)
[2] P Olver, Application of Lie groups to differential equations (Springer, New York, 1993)
[3] E Noether, Mathematisch-Physikalische Klasse 2, 235 (1918), English translation in Transport
Theory and Statistical Physics 1(3), 186 (1971)
[4] A H Bokhari and A H Kara, Gen. Relativ. Gravit. 39, 2053 (2007)
[5] A H Bokhari, A H Kara, A R Kashif and F D Zaman, Int. J. Theor. Phys. 45(6), 1029 (2006)
[6] A D Rendall, Applications of the theory of evolution equations to general relativity, Gen-
eral Relativity – Proceedings of the 16th International Conference edited by N T Bishop and
S D Maharaj (World Scientific, Singapore, 2002)
[7] Ahmad M Mugbil, On Noether symmetry and conservation laws, Ph.D. Thesis (King Fahd
University of Petroleum and Minerals, Dhahran, Saudi Arabia, 2011).
[8] C W Misner, K S Thorne and J A Wheeler, Gravitation (W H Freeman, San Francisco, 1973)
[9] G Bluman and S Kumei, Symmetries and differential equations (Springer-Verlag, New York,
1989)
[10] W Hereman, Int. J. Quantum Chem. 106, 278 (2006)
[11] A H Kara and F M Mahomed, Int. J. Theor. Phys. 39(1), 23 (2000)
[12] A H Kara and F M Mahomed, J. Nonlinear Math. Phys. 9, 60 (2002)
[13] U Göktas and W Hereman, Physica D 123, 425 (1998)