Ecuaciones Diferenciales Mediante Simetrías

Ecuaciones Diferenciales
Mediante Simetras
por
Jos Antonio Belinchn
Primer Borrador Junio 2005.
Segundo Borrador Septiembre 2006
ltimo Borrador Mayo 2009
ii
Contents
Prlogo v
1 Dimensional Analysis 1
1.1 Assumptions in DA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The Recipe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 The Pi-Theorem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Examples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Applications of dimensional analysis to ODEs and PDEs . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5.1 Differential equations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5.2 Partial differential equations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2 Introduction to the symmetry methods. 15
3 Ecuaciones de Primer Orden 25
3.1 Introduccin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2 The method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.3 Cuadraturas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.3.1 Simetras. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3.2 Ejemplos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4 Ecuaciones en variables separables y reducibles a ellas. . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4.1 Mtodo tradicional. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4.2 Mtodo dimensional. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4.3 Mtodo de Lie. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4.4 Ejemplo. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.5 Ecuacin homognea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
i
ii CONTENTS
3.5.3 Mtodo de Lie. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.5.4 Ejemplos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.6 Ecuacin lineal de primer orden . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.6.3 Mtodo de Lie. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.6.4 Ejemplos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.7 Ecuaciones exactas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.8 Ecuacin de Bernoulli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.8.3 Mtodo de Lie. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.8.4 Ejemplos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.9 Ecuaciones tipo Riccati. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.9.1 Ejemplos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.10 Ecuaciones tipo Abel y Chini. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.10.1 Mtodo tradicional . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.10.2 Ejemplos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.10.3 Pathological cases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.11 Ecuaciones de DAlambert-Lagrange y de Clairaut . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.11.3 Ejemplos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.12 Conclusions and discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4 Ecuaciones de segundo orden. 83
4.1 Introduccin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.2 Distintos tipos de ecuaciones. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.2.1 Lineales con coeciente constantes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.2.2 Ecuacin de Bessel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.2.3 Ecuacin de Dufng. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.2.4 Ecuacin de Emden y Emden-Fowler. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
CONTENTS iii
4.2.5 Ecuacin lineal exacta. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.2.6 Ecuacin no lineal exacta. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.2.7 La ecuacin lineal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.2.8 Ecuacin de Largerstrom. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.2.9 Ecaucin de Lienard. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.2.10 La ecuacin de Liouville. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.2.11 La ecuacin de Painleve. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.2.12 La ecuacin de Titchmarch. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.2.13 La ecuacin de Van der Pol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.3 Ejemplos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5 Ecuaciones de tercer orden. 113
5.1 Ejemplos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6 Aplicaciones en problemas cosmolgicos. 121
6.1 Distintas Aplicaciones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.1.1 Lie method for the H-equation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.1.2 Fluidos perfectos con constantes variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6.1.3 Fluidos viscosos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
iv CONTENTS
Prlogo
Estas notas sobre ecuaciones diferenciales ordinarias (odes) anlisis dimensional (AD) y grupos de Lie (GL) estn
divididas en varios captulos (no terminados) como sigue:
1. Cap.1. En este captulo (no terminado) dedicado al AD intento resaltar la estructura de grupo (grupo de
Lie, grupo generado por la simetra de escala) que tiene, con esto quiero decir que la solucin que uno
obtiene aplicando el teorema Pi (a una ode dimensionalmente homognea) es la misma que la generada
(obtenida) como solucin invariante por la simetra de escala. Esto es debido a la estructura de grupo de
tiene el AD. Para intentar hacer ver como funciona el mtodo expongo un ejemplo muy concreto de un
problema cosmolgico. Tal y como est redactado este captulo, pienso que uno debera intentar jarse
ms en el mtodo y no en el ejemplo.
Observacin 0.0.1 El AD que sigo es bsicamente el AD de J. Palacios aunque he tomado (adoptado) todas las
aportaciones y correcciones que M. Castas ha llevado a cabo para superar los postulados y mtodos de Palacios.
2. Cap. 2. Estar dedicado a exponer (sin demostrar) el mtodo de los grupos de Lie a las odes. Espero
terminar de redactar estos dos primero captulos durante el otoo de 2004. Todas las referencias de este
captulo tambin son vlidas para los captulos 3 y 4.
Creo y espero que los captulos interesantes (en su forma actual) son los captulos 3 y 4.
3. Cap. 3. En este captulo expongo algunas de las odes de primer orden (mediante ejemplos) y comparo su
resolucin por tres mtodos:
El tradicional, en el que uno debe clasicar (saber de que tipo se trata, identicacin previa) y luego
aplicar el cambio de variable pertinente (apropiado) i.e. un rollo pues por ejemplo con odes tipo
Riccati adems debemos buscar soluciones particulares, en n lo de siempre.
El dimensional. Para aplicar este mtodo lo primero que tenemos que hacer (comprobar) es que la ode
bajo estudio verique el principio de homogeneidad dimensional i.e. de forma pedestre se puede
decir que: peras = peras pero nunca peras = manzanas, para que se verique la ltima igual deber-
emos introducir ciertas constantes dimensionales que hagan que dicha ode verique el principio de
homogeneidad i.e.
y
= x y
= ax / [a] = x
2
y
(ecuaciones con sentido fsico).
He seguido dos tcticas dimensionales:
D1. (Mtodo pedestre) Consiste en aplicar el AD (Teorema Pi) a la ode bajo estudio y obtener (intentar
obtener) una solucin. Esta solucin por lo general ser una solucin particular de la ode dada pero
adems ser una solucin invariante y atractiva desde el punto de vista de los sistemas dinmicos
y con sentido fsico (esto si es bastante importante). Esta tctica no siempre funciona y es bastante
limitada.
D2. Consiste en obtener variables nuevas mediante el teorema Pi y reformular (reescribir) la ode dada
en estas nuevas variables, la ode resultante es trivialmente integrable, este es en esencia el mtodo de
v
vi CONTENTS
Lie pero muchsimo ms sencillo de aplicar ya que uno no debe aplicar ningn algoritmo para obtener
las simetras que inducen dichos cambios de variable. Esta tctica siempre funciona.
Con estas dos tcticas uno puede olvidarse de clasicaciones y del correspondiente cambio de variable
qu eno es poca cosa.
El mtodo de Lie. Consiste en buscar mediante un algoritmo (algoritmo de Lie) las posibles simetras de
la ode. Estas simetras inducen cambios de variables que nos permiten reescribir la ode bajo estudio
obteniendo as una ode que admite una integracin trivial. El estudiar, aplicar, las odes de primer
orden mediante este mtodo es ir a matar moscas a caonazos, es casi ms complicado aplicar el
mtodo (resolver la edp resultante al aplicar el algoritmo de Lie) que acordarse del pertinente cambio
de variable, pero no en todos los casos, por ejemplo con las odes tipo Riccati nos evita el tener que
buscar una solucin particular y en los casos ms favorables nos permite encontrar una solucin a las
odes tipo Abel o a las de tipo Chini.
He intentado resaltar que la solucin invariante que induce la simetra de escala coincide con la
obtenida al aplicar el teorema Pi (nuestro mtodo pedestre D1).
4. Cap. 4. En este captulo repaso algunas de las odes de segundo orden, (tampoco est del todo terminado).
Mediante ejemplos intento hacer ver como funcionan tanto el mtodo de Lie (bsicamente) como el di-
mensional. Es aqu donde parece tener ms utilidad el mtodo de Lie. En el quinto captulo estudio slo
algunas ODEs de tercer orden.
En el ltimo captulo expongo unos ejemplos de aplicacin del mtodo de Lie a ecuacines de segundo
orden que me han surgido en el estudio de modelos cosmolgicos. Aqu la idea es la utilizar los GL y
el AD no slo para obtener soluciones a la ode dada sino para estudiar condiciones que deben cumplir
las funciones involucradas para que dicha ode sea integrable, por ejemplo expongo una ode de segundo
orden en la que estn involucradas 3 funciones, entonces el mtodo de Lie nos dice que condiciones deben
cumplir para que dicha ode sea integrable. De igual forma nos ayuda a establecer ecuaciones de estado
(casi nada).
Lamento de antemano todas las erratas, faltas de ortografa etc... y espero que estas notas puedan ser tiles para
alguien.
CONTENTS vii
viii CONTENTS
Chapter 1
Dimensional Analysis
1.1 Assumptions in DA.
DA consider the following concepts:
(i) A quantity u is to be determined in terms of n measurable quantities (variables and constants, characteris-
tic and universal) (W
i
)
n
i=1
;
u = f (W
1
, ..., W
n
) , (1.1)
where f is a unknown function of (W
i
)
n
i=1
.
(ii) The quantities
_
u; (W
i
)
n
i=1
are measured in term of m fundamental dimensions labeled as dimensional

base B = {
_
L
j
_
m
j=1
}. We will put special emphasis in calculating the multiplicity of this dimensional base
i.e. the number of fundamental dimensions. For example in the case in which we may take into account
the spatial discrimination, see the section of examples.
(iii) With respect to this dimensional base a quantity Z has the following dimensional equation, i.e. has the
following dimensions with respect to this base, denoted as:
[Z] = L
1
1
.....L
m
m
, (1.2)
where the numbers (
i
)
n
i=1
, with
i
Ri = 1, ..m, are the dimensional exponents of Z.
As we will see through the examples, the choice of the dimensional base will be essential in order to obtain
good results applying DA to diverse physical or mathematical problems.
(iv) We say that a quantity Z is dimensionless (with respect to B) written
[Z] = 1, (1.3)
iff (
i
)
n
i=1
= 0, for all
i
Ri = 1, ..m.
(v) Dimensional homogeneity of the equations under study. One we have xed our dimensional base we
need to check if our equations verify the principle of dimensional homogeneity with respect to this di-
mensional base, otherwise we will need to introduce the appropriate dimensional constants that make
homogeneous our equations under study.
(vi) For any set of fundamental dimensions, one can choose a system of units for measuring the value of any
quantity Z, for example the usual International System of units. A change from one system of units to
another involves a positive scaling of each fundamental dimension that in turn induces a scaling of each
quantity Z. Under a change of system of units, the value of a dimensionless quantity is unchanged i.e. its
value is invariant under an arbitrary scaling of any fundamental dimension. Therefore, it is meaningful
1
2 CHAPTER 1. DIMENSIONAL ANALYSIS
to deem dimensionless quantities as large or small. The last assumption of D.A. is that formula (1.1) acts
as a dimensionless equation in the sense that (1.1) is invariant under arbitrary scaling of any fundamental
dimension i.e. it is independent of the choice of a system of units.
Next, we see a particular example in order to clarify all these abstract ideas. For this purpose we may consider
the well known law of Newton
F = ma = m
d
2
x
dt
2
, (1.4)
where we are interested in calculating the velocity, v, of some body, so v is the quantity to calculate. If we rewrite
the above equation (1.4) as
Fm
1
x
1
t
2
= 1, (1.5)
we may choose as dimensional base the following one B = {x, m, t} that we rewrite as
B = {L, M, T} , (1.6)
where L stands for length, M for mass and T for time, where the elements belonging to the dimensional base
are usually written in capital letters instead of small letters. Note that we have only one equation and four
quantities, hence the multiplicity of the dimensional base is: (number of quantities)(number of equations), in fact
is the rank of the dimensional matrix (see below, the recipe, for clarifying this calculation). In the next chapter
we will clarify this point.
In this way we say that the dimensional equation of the force F is
[F] = LMT
2
, (1.7)
while the dimensional equation of the quantity v is
[v] = LT
1
. (1.8)
Other bases can be obtained, but I am not interested in exploring this fact. We only stress that it is very important
calculate the appropriate base for each problem as we will see in the following sections (see section of examples).
Now we may check if our equation veries the principle of dimensional homogeneity, in such a way that
[F] = [m] [a] , LMT
2
= M LT
2
, (1.9)
since [a] = LT
2
.
This example is trivial, but we are only trying to clarify these new concepts.
1.2 The Recipe.
The above concepts are the main assumptions of the Pi-Theorem. Pi-Theorem allows us to calculate the relation-
ship between the quantity u and the set of quantities (W
i
)
n
i=1
, the set of variables and constants, characteristic
and universal. These assumptions bring us to the following conclusions:
1. Formula (1.1) can be expressed in term of dimensionless quantities.
2. The number of dimensionless quantities is
k + 1 = n + 1 rg(B) (1.10)
where rg(B) is the rank of the dimensional matrix. This matrix is formed by the column vectors
b
i
=
_
_
_
_
_
b
1i
b
2i
.
.
.
b
mi
_
_
_
_
_
= B =
_
_
_
_
_
b
11
b
12
b
1n
b
21
b
22
b
2n
.
.
.
.
.
.
.
.
.
.
.
.
b
m1
b
m2
b
mn
_
_
_
_
_
(1.11)
1.3. THE PI-THEOREM. 3
where b
i
is the dimension vector of W
i
, i = 1, ..., n.
Precisely k of these dimensionless quantities depend on the measurable quantities (W
i
)
n
i=1
.
3. Let
x
(i)
=
_
_
x
1i
x
2i
.
.
.
x
mi
_
_
, (1.12)
with i = 1, ..., k, represent the k = n rg(B) linearly independent solutions x of the system Bx = 0. Let
a =
_
_
a
1
a
2
.
.
.
a
m
_
_
, (1.13)
be the dimension vector of the quantity u, and let
y =
_
_
y
1
y
2
.
.
.
y
n
_
_
, (1.14)
represent a solution of the system By = a. Then formula (1.1) simplies to
= g (
1
, ....,
k
) , (1.15)
where
_
, (
i
)
k
i=1
_
are dimensionless quantities given by
= uW
y
1
1
W
y
2
2
....W
y
n
n
,
i
= W
x
1i
1
W
x
2i
2
....W
x
ni
n
, (1.16)
with i = 1, ...k, and g is some unknown function of its arguments. In particular eq. (1.1) becomes
u = W
y
1
1
W
y
2
2
....W
y
n
n
g (
1
, ....,
k
) . (1.17)
1.3 The Pi-Theorem.
Now we will give a brief sketch of the proof of this famous theorem. First of all
[u] = L
a
1
1
.....L
a
m
m
, [W
i
] = L
b
1i
1
.....L
b
mi
m
, (1.18)
with i = 1, ...n. Next, we use assumption (vi) and consider the invariance of eq. (1.1) under arbitrary scaling of
the fundamental dimensions by taking each fundamental dimensions in turn. We rst scale L
1
by letting
L
1
= e
L
1
, R, (1.19)
in turn this induces the following scalings of the measurable quantities
u
= e
a
1
u, W
i
= e
b
1i
W
i
, i = 1, ...n. (1.20)
Equations (1.20) dene a one-parameter () Lie group of scaling transformations of the n+1 quantities (u, W
1
, W
2
, ..., W
n
)
with = 0 corresponding to the identity transformation. This group is induced by the one-parameter group of
scalings (1.19) of the fundamental dimension L
1
.
From the assumption (vi), eq. (1.1) holds iff
u
= f (W
1
W
2
.....W
n
) , (1.21)
i.e.
e
a
1
u = f
__
e
b
1i
W
i
_
n
i=1
_
, R, (1.22)
then two cases need to be distinguished:
Case I. b
11
= .... = b
1n
= 0 = a
1
. Here L
1
, is not a fundamental dimension of the problem, i.e. eq. (1.1) is
dimensionless with respect to L
1
.
Case II. b
11
= .... = b
1n
= 0, and a
1
= 0. Hence u = 0 is a trivial solution.
Therefore it follows that b
1i
= 0 for some i = 1, ...n. Without loss of generality we assume b
11
= 0. We dene
new measurable quantities
X
i1
= W
i
W
b
1i
/b
11
1
, i = 2, 3, ...n, (1.23)
and let
X
n
= W
1
. (1.24)
We choose as the new unknown
v = uW
a
1
/b
11
1
, (1.25)
the transformation given by eqs. (1.23)-(1.25) denes a one-to-one mapping of the quantities (W
i
)
n
i=1

(X
i
)
n
i=1
, and a one-to-one mapping of the quantities
_
u, (W
i
)
n
i=1

_
v, (X
i
)
n
i=1
, in this way eq. (1.1) is equiv-

alent to
v = F
_
(X
i
)
n
i=1
_
, (1.26)
where F is a unknown function of (X
i
)
n
i=1
. Thus the group of transformations (1.20) becomes
v
= v, (1.27)
X
i
= X
i
, i = 1, 2, ....n 1, (1.28)
X
n
= e
b
11
X
n
, (1.29)
therefore
_
v, (X
i
)
n1
i=1
_
are invariants of (1.20). Moreover, the quantities
_
v, (X
i
)
n
i=1
satisfy assumptions (iii), and

eq. (1.26) satises assumption (vi). Therefore
v = F
_
(X
i
)
n1
i=1
, e
b
11
X
n
_
, R. (1.30)
where F is independent of the quantity X
n
. Moreover, the measurable quantities (X
i
)
n1
i=1
are products of powers
of (W
i
)
n
i=1
and v is a product of u and powers of (W
i
)
n
i=1
. Thus eq. (1.1) reduces to
v = G
_
(X
i
)
n1
i=1
_
, (1.31)
where
_
v; (X
i
)
n1
i=1
_
are dimensionless with respect to L
1
and G is some function of (n 1) arguments.
If we repeat this tactic with the (m1) fundamental dimensions, we reduce eq. (1.1) to a dimensionless formula
= g
_
(
i
)
k
i=1
_
, (1.32)
where [] = [
i
] = 1, with i = 1, ..., k, and g is some function of (
i
)
k
i=1
, being
= u
n
i=1
W
y
i
i
,
i
=
k
j=1
W
x
mj
m
, (1.33)
with m = 1, ..., n.
1.4. EXAMPLES. 5
Note that this theorem makes no assumptions about the continuity of the function f and hence of g with respect
of any of their arguments.
It is shown immediately that the number of measurable dimensionless quantities is k = n rg(B). This follows
since
_
W
x
1
1
W
x
2
2
....W
x
n
n
= 1, (1.34)
iff
x =
_
_
x
1
x
2
.
.
.
x
n
_
_
, (1.35)
satises Bx = 0. This equation has k = n rg(B) linearly independent solutions x
(i)
given by (1.12). The real
numbers (1.14) follow from setting
_
uW
y
1
1
W
y
2
2
....W
y
n
n
_
= 1, (1.36)
leading to y satisfying By = a.
1.4 Examples.
In this section we will try to explain how works DA through different examples and how we can avoid the cases
which we obtain a result with an unknown function g. For this purpose we will follow two different tactics,
the rst of them will consist in making a mathematical hypothesis on the behaviour of this unknown function
and we will justify (as far as possible) such assumption. In the second of our tactics we will use the so called
dimensional discrimination. Such tactic allows us to obtain a complete solution for our problems and it is not
needed to make any previous assumption.
Ejemplo 1.4.1 The famous atomic explosion of 1945.
Solution. We are interested in knowing the radius R of the shock wave that is produced an atomic explosion.
For this purpose we treat R as the unknown (u = R with the previous notation) and assume that
R = f (W
1
, W
2
, W
3
, W
4
), (1.37)
where W
1
= E is the energy released by the explosion, W
2
=
0
is the ambient air density, W
3
= P
0
is the ambient
air pressure, W
4
= t is the elapsed time after the explosion takes place. For this problem we use B = {L, M, T}
as dimensional base. therefore the corresponding dimensional matrix is given by
E t P
0

0
L 2 0 1 3
M 1 0 1 1
T 2 1 2 0
, / B =
_
_
2 0 1 3
1 0 1 1
2 1 2 0
_
_
(1.38)
since for example [E] = L
2
MT
2
, etc... and as we can see the rg(B) = 3 therefore k = n rg(B) = 4 3 = 1 is
the number of monomia.
Following the above explained procedure we nd that Bx = 0 i.e.
_
_
2 0 1 3
1 0 1 1
2 1 2 0
_
_
_
_
_
_
x
1
x
2
x
3
x
4
_
_
_
_
= 0, =
_
_
_
x
1
=
2
5
x
4
x
2
=
6
5
x
4
x
3
=
3
5
x
4
, (1.39)
where x
4
is arbitrary. Setting x
4
= 1 we obtain the measurable dimensionless quantity
1
= P
0
_
t
6
E
2
3
0
_
1/5
. (1.40)
Now the dimension vector of R is
a =
_
_
1
0
0
_
_
, [R] = L, (1.41)
the general solution of By = a is
y =
1
5
_
_
_
_
1
2
1
0
_
_
_
_
+ x, (1.42)
where x is the general solution of Bx = 0. Setting x = 0, we obtain the dimensionless unknown
= R
_
Et
2
0
_
1/5
, (1.43)
therefore the solution that DA gives us is the following one
R =
_
Et
2
0
_
1/5
g(
1
), (1.44)
where g is some function of
1
.
How to solve the big drawback of having an unknown function g?
We may make the following assumption (called the power law hypothesis, from now on)
= (
1
)
n
, n R, (1.45)
i.e. we are supposing that the general solution follows a power law and we x the value of the numerical
constant n through experiment. Therefore the solution that we suggest take the following form
R =
_
Et
2
0
_
1/5
_
_
P
0
_
t
6
E
2
3
0
_
1/5
_
_
n
= E
12n
5
t
2+6n
5

3n1
5
0
P
n
0
, (1.46)
basically we are assuming that
R Ct
2+6n
5
near t = 0. (1.47)
where C is a constant.
Numerical values from the experiments (i.e. after many nuclear explosions) bring us to obtain n = 0, so the
solution is
R =
_
E
0
_1
5
t
2
5
= R = At
2/5
. (1.48)
Observacin 1.4.1 The short cut. The problem is easily reduced to:
R = f (E, t, P
0
,
0
) , B = {L, M, T} (1.49)
where the dimensional matrix is given by
R E t P
0

0
L 1 2 0 1 3
M 0 1 0 1 1
T 0 2 1 2 0
. (1.50)
1.4. EXAMPLES. 7
Since the number of quantities is 5 and the multiplicity of the dimensional base is 3 therefore we will obtain 2
monomials, i.e. a solution like = g(
1
) where = f (R; E, t,
0
) and
1
= f
1
(E, t, P
0
,
0
).
From the dimensional matrix (1.50) we see that
1 + 2
E
P
0
3
0
= 0, (1.51)
E
+
P
0
+
0
= 0, (1.52)
2
E
+
t
2
P
0
= 0 (1.53)
therefore the solution under the assumption = (
1
)
n
is
R = E
E
t
0
0
P
P
0
0
, i.e. R = E
12n
5
t
2+6n
5

3n1
5
0
P
n
0
(1.54)
as we can deduce in a trivial way from the above system of linear equations (1.51)-(1.52).
We need to avoid this kind of assumptions, although they work well, or at least to try to justify mathematically why we may
expect this class of solutions i.e. power law solutions.
Ejemplo 1.4.2 The range of a projectile.
A projectile of mass m weight W and with initial velocity v, is launched in horizontal direction from a height h. We are
interested in knowing the range x of the projectile.
Solution. In this case we will explore two possibilities, in the rst of them the application of DA brings us to
obtain an unsatisfactory solution since this depends on an unknown function. To solve this drawback we explore
another possibility which consists in enlarge the dimensional base taking into account the spacial discrimination.
1. Briey the problem is reduced to:
x = f (m, W, v, h) , (1.55)
that with respect to the dimensional base B = {L, M, T} it is obtained the following dimensional matrix
x m W v h
L 1 0 1 1 1
M 0 1 1 0 0
T 0 0 2 1 0
, = =
x
h
,
1
=
mv
2
hW
(1.56)
therefore the solution will be
= (
2
), i.e. x = h
_
mv
2
hW
_
. (1.57)
As we have already commented this solution is very poor for a very simple problem, for this reason we
will try another tactic.
2. As we can see there are two privileged spatial directions in this problem
_
L
x
, L
y
_
. For this reason, under
this new consideration, the problem is reduced to (as above):
x = f (m, W, v, h) , (1.58)
but with respect to the dimensional base B =
_
L
x
, L
y
, M, T
_
it is obtained the following dimensional
matrix
x m W v h
L
x
1 0 0 1 0
L
y
0 0 1 0 1
M 0 1 1 0 0
T 0 0 2 1 0
, (1.59)
which bring us to obtain only one monomia. Therefore the solution is in this case
x = v
_
hm
W
= v
h
g
, (1.60)
as it is expected.
With this example we have tried to show another tactic to solve physical problem and without the necessity of
doing assumptions on the behaviour of the solution as in the previous example.
Ejemplo 1.4.3 The golfer and the moral. We are interested in knowing how long, t, must train a golfer of mass m and
weight W, in order to hit a golf-ball to a distance l.
Solution. As in the above example we may see that the set of governing parameters is
P = P (t; m, W, l) , (1.61)
that with respect to the dimensional base B = {L, M, T}, we obtain the following the dimensional matrix
t m W l
L 0 0 1 1
M 0 1 1 0
T 1 0 2 0
, = t = C
l
g
, (1.62)
from which we obtain a single monomial, as it is observed in a trivial way. C R.
Observacin 1.4.2 Obviously this solution is a complete nonsense since, from our point of view, it is necessary that the
problem must be considered within some physical theory (and this is not the case). DA is subordinate to one physical
theory and not the other way round as we may deduce from this simple example although some times one may found some
interesting relationship between the quantities i.e. with physical meaning.
1.5 Applications of dimensional analysis to ODEs and PDEs
In this section we will try to show how to apply D.A. to solve differential equations and partial differential
equations.
1.5.1 Differential equations.
The method of the Lie groups it has showed as a very useful tool in order to solve nonlinear equations (nl-ODE)
as well as PDE. Nevertheless, when one is studying ode of rst order its application becomes very complicated
(very tedious) if one has not a computer algebra package since if one decides to look for the possible symmetries
of an ODE with pencil and paper this task may be turned very exhausting. However, if we know that an ODE
admits a concrete symmetry, then it is a trivial issue to nd new variables which allows us to rewrite the ODE in
quadratures or as we will see in this paper to obtain an ode with separating variables.
1.5. APPLICATIONS OF DIMENSIONAL ANALYSIS TO ODES AND PDES 9
Our purpose in this work is to explain, through an example, how D.A. works in order to nd these changes
of variables (c.v.) in a trivial way, i.e. without the knowledge of the symmetries of the ODE under study. The
idea is as follows. When we are studying an ODE from the dimensional point of view, we must require that
such ODE veries the principle of dimensional homogeneity (pdh) i.e. that each term within the equation have
the same dimensions, for example, speed, or energy density. To clarify this concept we consider the following
ode, y
=
y
x
+ x, where each term must have dimensions of y
i.e. [y
] = yx
1
, where [] stands for dimensional
equation of the quantity . As we can see, this ODE does not verify the pdh, since the term x has dimensions of
x, i.e. [x] = x. In order to do that this equation veries the pdh we need to introduce dimensional constants,
in such a way that after rewrite the ODE with these constants the ODE now veries such principle i.e. in this
example and as we can see easily if we consider only one constant a, such that [a] = yx
2
, we make that the
ODE y
=
y
x
+ ax, veries the pdh. We would like to emphasize that this situation does not arise when one is
studying physical or engineering problems since, as it is supposed, such problems (equations) verify the pdh
and we do not need to introduce new dimensional constants, that must have physical meaning, for example, the
viscosity coefcient, etc....
Precisely this dimensional constant suggests us the c.v. (x, y(x)) (t, u(t)) where
_
t = x, u = yx
2
_
in such
a way that rewriting the original ODE in these new variables yields a new ODE with separating variables:
u + u
t = 1. The reason is the following. We know from the Lie group theory that if it is known a symmetry
of an ODE then this symmetry bring us through a c.v. to obtain a simpler ODE (a quadrature or a ODE with
separating variables) and therefore the solution is found in a closed form. To nd this c.v. one uses the invariants
that generate each symmetry, i.e. the rst principle is that it is useful to pass to new coordinates such that one
of the coordinate functions is an invariant of the group. After such transformation it often (but not always)
happens that the variables separate and the equation can be solved in closed form. We must stress that taking
the new dependent variables to be an invariant of the group does not guarantee the separation of variables. The
choice of the independent variable is also very important.
One must note that D.A. gives us this invariant (or at least a particular solution). It is observed in our example
that the solution y = x
2
, (suggested by D.A.) is a particular solution for this ODE, but if we study this ode form
the Lie theory point of view it is obtained that such ODE has the scaling symmetry X = x
x
+2y
y
and therefore
the solution y = x
2
, is furthermore an invariant solution (generated by X), for this reason such c.v. works well.
As we will show in the following example, the c.v. that D.A. induces is the same as the one generated by the
symmetries of the ODE. In order to make see this fact, we will solve the example by D.A. as well as by the Lie
group technique, calculating the symmetries and their corresponding invariants and c.v..
Ejemplo 1.5.1 Solve the Abel ODE.
y
= Cx
3
y
3
+ Bxy
2
A
y
x
, (1.63)
Solution. If we rewrite eq. (3.256) introducing the following dimensional constants,
y
= a
2
Cx
3
y
3
+ aBxy
2
A
y
x
, (1.64)
where A, B, C R, and [a] = X
2
Y
1
. As we can see this ODE is scale invariant since we have needed to
introduce only one constant. DA suggests us the following c.v.
_
t = x, u(t) = ax
2
y
_
=
_
x = t, y =
u
at
2
_
, (1.65)
in such a way that eq. (3.256) is written in the following form:
tu
= u
_
u
2
+ u + 1
_
, (1.66)
and its solution is:
ln t +
1
2
ln
_
u
2
+ u + 1
_
+
3
3
arctan
__
3
2
u +
1
3
_
3
_
ln u + C
1
= 0, (1.67)
therefore in the original variables it yields:
ln x +
1
2
ln
_
_
ax
2
y
_
2
+ ax
2
y + 1
_
+
3
3
arctan
__
3
2
_
ax
2
y
_
+
1
3
_
3
_
ln
_
ax
2
y
_
+ C
1
= 0. (1.68)
In second place, we study eq. (3.256)
y
= Cx
3
y
3
+ Bxy
2
A
y
x
, (1.69)
with respect to the dimensional base B = {T} . This ODE veries the principle of dimensional homogeneity with
respect to this dimensional base. Note that [y] =
_
1
H
_
= T
2
, and [x] = [H] = T
1
hence [y
] = T
3
. Therefore
rewriting the equation in a dimensionless way we nd that y x
2
. This ODE has been obtained studying
a cosmological model and therefore H is the usual Hubble parameter (see the next chapter to understand the
meaning of this quantity) and therefore [H] = T
1
as it is usual and T stands for dimensions of time.
But if we study this equation with respect to the dimensional base B = {X, Y} , we need to introduce new
dimensional constants that make the equation verify the principle of dimensional homogeneity
y
= Cx
3
y
3
+ Bxy
2
A
y
x
, (1.70)
where
_
1/2
_
= [] = X
2
Y
1
, hence
y x
X 0 2 1
Y 1 1 0
= y
1
x
2
, (1.71)
which is a particular solution of the original ode.
Lie Method. In order to nd the symmetry generator, X = (x, y)
x
+ (x, y)
y
, of a ODE
y
= f (x, y) , (1.72)
we need to solve the following PDE
x
+ (
y
x
) f
y
f
2
(x, y) f
x
(x, y) f
y
= 0, (1.73)
where
f
x
=
d f
dx
, f
y
=
d f
dy
. (1.74)
see the appendix in order to clarify notation and concepts.
In this case we have to solve
x
+ (
y
x
)
_
Cx
3
y
3
+ Bxy
2
A
y
x
_
y
_
Cx
3
y
3
+ Bxy
2
A
y
x
_
2
(x, y)
_
3Cx
2
y
3
+ By
2
+ A
y
x
2
_
(x, y)
_
3Cx
3
y
2
+ Bxy
A
x
_
= 0, (1.75)
This ODE admits the following symmetry
X = x
x
2y
y
, (1.76)
which is a scaling symmetry and it induces the following change of variables,
Xr = 0, Xs = 1, (1.77)
r = x
2
y, s(r) = ln(x), = x = e
s(r)
, y =
r
e
2s(r)
, (1.78)
which brings us to obtain the next ode in quadratures
s
=
1
r (Cr
2
+ Br + 2 A)
, (1.79)
and whose solution is:
s(r) =
lnr
A 2
+
1
2
ln
_
Cr
2
+ Br + 2 A
_
A 2

Barctanh
_
2Cr+B
B
2
+4C(A2)
_
(A 2)
_
B
2
+ 4C(A 2)
+ C
1
, (1.80)
and hence in the original variables (x, y):
ln x =
ln
_
x
2
y
_
A 2
+
1
2
ln
_
Cx
4
y
2
+ Bx
2
y + 2 A
_
A 2

Barctanh
_
2Cx
2
y+B
B
2
+4C(A2)
_
(A 2)
_
B
2
+ 4C(A 2)
+ C
1
, (1.81)
which is the most general solution for this ODE.
Now if we consider the invariants I that induce the symmetry X i.e.
dx
=
dy
, y
:=
dy
dx
=

, (1.82)
Hence, for example the symmetry X = x
x
2y
y
generates the following invariant:
dx
=
dy
=
dx
x
=
dy
2y
= ln x =
1
2
ln y = I = x
2
y y = ax
2
, (1.83)
this would be the solution that suggests us precisely the direct use of the Pi theorem(if the ODE is scale invariant,
as in this case, the solution obtained applying by the Pi theorem coincides with the invariant solution that
generates the scale symmetry, in this case X). We see that we only obtain a particular solution, but that this is
invariant, in fact if we think about the ODE as a dynamical system we see that the xed point of such equation
would be precisely the solution y = Kx
2
.
Observacin 1.5.1 Usually the most general solution of this kind of ODEs is unphysical i.e. it lacks of physical meaning.
So, why the invariant solution has physical meaning and is stable from the dynamical point of view?
1.5.2 Partial differential equations.
Using the notation of the section 2, now we consider a boundary value problem for a partial differential equation
(PDE)
_
u, (W
i
)
n
i=1
_
where u is the unknown function (the dependent variable of the PDE) and (W
i
)
n
i=1
are the
independent variables and constants. From the Pi-theorem it follows that the PDE may be re-expressed in
dimensionless form
_
, (
i
)
k
i=1
_
, being the dimensionless dependent variable and (
i
)
k
i=1
the dimensionless
independent variables and dimensionless constants.
Considering the set of quantities (W
i
)
n
i=1
we distinguish between (W
i
)
l
i=1
variables and (W
i
)
n
i=l+1
constants. Let
B be the dimensional matrix such that
B = (B
1
| B
2
) , (1.84)
of all (W
i
)
n
i=1
and B
1
= M((W
i
)
l
i=1
) the dimensional matrix of the quantities and B
2
= M((W
i
)
n
i=l+1
) the
dimensional matrix of the constants.
An important objective in applying D.A. to PDE is to reduce the number of independent variables. The rank
of B
2
i.e. r(B
2
) represents the reduction in the number of constants through DA. Consequently the reduction in
the number of independent variables is = r(B) r(B
2
). In particular the number of dimensionless measur-
able quantities is k = n r(B) = [(l ) + (n l r(B
2
))] , where (l ) of the quantities are dimensionless
independent variables and (n l r(B
2
)) are dimensionless constants.
If = r(B) r(B
2
) = 0, then DA reduces the given boundary value problem to a dimensionless boundary
problem with (n l r(B
2
)) dimensionless constants. In this case the number of independent variables is not
reduced. If l 2, (l ) = 1, then the resulting solution of the boundary value problem is called self-similar.
Ejemplo 1.5.2 Stokes rst problem. Let us consider a motionless uid in the proximity of a plane wall. If the wall starts
moving abruptly along its own plane (starting current) with a constant speed U
0
, the Navier-Stokes equation reduce to:
u
t
=
2
u
y
2
. (1.85)
Find the velocity distribution of the uid as function of y, the distance to the plane as well as function of, t :i.e.
u = u (y, t)
Solution. Let u be the uid velocity in the direction of U
0
(xdirection), in this case the Navier-Stokes equations
reduce to the viscous diffusion equation
u
t
=
2
u
y
2
, (1.86)
with the boundary conditions:
u(y, 0) = 0, y > 0, (1.87)
u(, t) = 0, and u(0, t) = U
0
, t > 0 (1.88)
where is the kinematic viscosity.
We are going to explore two dimensional tactics, the rst of them is the usual one while the second will consist
in the spatial discrimination.
1. In this case the dimensional base isB = {L, T} and therefore we obtain the following dimensional matrix:
u U
0
y t
L 1 1 1 0 2
T 1 1 0 1 1
(1.89)
obtaining in this way 3 monomials and therefore we do not reduce the number of variables in the problem.
We have insisted in outlining this tactic in order to show the useful of the spatial discrimination as follows.
2. In this case the dimensional base is B =
_
L
x
, L
y
, T
_
and therefore the new dimensional equations of each
quantity are:
[] = L
2
y
T
1
, [u] = [U
0
] = L
x
T
1
, [y] = L
y
, [t] = T (1.90)
The most general solution is:
f (y, t, u, U
0
, ) = 0, (1.91)
we have the following dimensional matrix:
u U
0
y t
L
x
1 1 0 0 0
L
y
0 0 1 0 2
T 0 0 0 1 1
, (1.92)
obtaining two monomia :
=
u
U
0
,
1
=
y
t
, (1.93)
then, new variables and U are dened as
=
y
t
, U () =
u
U
0
, i.e. u = U
0
U () , (1.94)
Now if we calculate
u
t
and

2
u
y
2
as functions of U and , we obtain
d
2
U
d
2
+

2
dU
d
= 0 (1.95)
with the boundary conditions:
U(0) = 1, U() = 0, (1.96)
where U = 0 for t = 0, y > 0 (as, ) and U = 1 for t , y 0 (as, = 0). Therefore the solution is:
U = erf
_
2
_
. (1.97)
For example, the change of variables
U
= z, = s, (1.98)
brings us to obtain the following ode
z
=
1
2
sz = z = C
1
e
s
2
/4
, (1.99)
and making the inverse change of variables
U
= C
1
e
2
/4
, = U = C
1
_
e
2
/4
dx = erf
_
2
_
, (1.100)
it is obtained the solution (1.97).
In this way we can see how works the dimensional discrimination, showing us that is it a powerful tool to solve
partial differential equations.
Chapter 2
Introduction to the symmetry methods.
What are the symmetries of an ODE?. Consider the simplest example
y
=
dy
dx
= 0, (2.1)
whose solution is obviously the set of straight lines y = c, with c R. The ODE (2.1) is represented geometrically
by the set of all solutions and so any symmetry of the ODE must necessarily map the solution set into it self.
Formally, the condition that any symmetry maps the object into itself requires that the set of solutions curves in
the (x, y) plane must be indistinguishable from the image in the ( x, y) plane, so
y
= 0 when y
= 0. (2.2)
A smooth transformation of the plane is invertible if its Jacobian is nonzero so we impose the further condition
x
x
y
y
x
y
y
x
= 0. (2.3)
A particular solution curve will be mapped to a (possibly different) solution curve, and so
y (x, c) = c(c), c R. (2.4)
Here x is regarded as a function of x and c that is obtained by inverting x = x(x, c). ODE (2.1) is very simple and
so all of its symmetries can be found. Differentiating (2.4) with respect to x, we obtain y
x
(x, c) = 0, c R,
therefore taking (2.3) into account, the symmetries of (2.1) are of the form
( x, y) = ( f (x, y), g(y)) , f
x
= 0, g
y
= 0. (2.5)
where f and g are assumed to be smooth functions of their arguments. The ODE has very large family of
symmetries. We were able to use the know general solution of (2.1) to derive (2.2) which led to (2.5). However,
we could also have obtained this result directly from (2.2) as follows.
On the solution curves, y is a function of x, and hence x = (x, y) and y = (x, y) may be regarded as a function of
x. Then by the chain rule (2.2) can be written as
d y
d x
=
D
x
y
D
x
x
= 0, when y
= 0, (2.6)
where D
x
denotes the total derivative with respect to x, i.e. D
x
=
x
+ y
y
+ y
y
+ ..., therefore (2.2) amounts
to
y
x
+ y
y
y
x
x
+ y
x
y
= 0, when y
= 0, (2.7)
and therefore (2.5) holds.
15
16 CHAPTER 2. INTRODUCTION TO THE SYMMETRY METHODS.
The advantage of using the symmetry condition in the form (2.2) is that one can obtain information about the
symmetries without having to know the solution of the differential equation in advance.
For an ODE
y
= (x, y), (2.8)

the symmetry condition reads
y
= ( x, y), when y
= (x, y). (2.9)

As before, we regard y as function of x on the solution curves, therefore (2.9) yields
D
x
y
D
x
x
=
y
x
+ y
y
y
x
x
+ y
x
y
= ( x, y), when y
= (x, y), (2.10)

therefore the symmetry condition for (2.8) is equivalent to the constraint
y
x
+ (x, y) y
y
x
x
+ (x, y) x
y
= ( x, y), (2.11)
together with the requirement that the mapping should be a diffeomorphism.
Now, suppose that we are able to nd a nontrivial one-parameter Lie group (LG) of symmetries of the ODE (2.8)
then the Lie group can be used to determine the general solution of the ODE. Suppose that the symmetry of
(2.8) include the Lie group of transformations in the y-direction i.e. (0, 1)
( x, y) = (x, y + ) , (2.12)
then, the symmetry condition (2.11) reduces to
(x, y) = (x, y + ) , (2.13)
for a very small values of . Differentiating (2.13) with respect to at = 0,
y
(x, y) = 0, and therefore the most
general ODE whose symmetries include the Lie group (2.12) is of the form
y
= (x), = y =
_
(x)dx + c, (2.14)
the articular solution corresponding to c = 0 is mapped by the translation to the solutions
y =
_
(x)dx + =
_
( x)d x + , (2.15)
which is the solution corresponding to c = . So by using the one-parameter LG we obtain the general solution
from one particular solution.
Is the same true for ODEs with other one-parameter LG?. Yes, the same idea works for all one-parameter LG.
In a suitable coordinate system, the symmetries parametrized by sufciently close to zero are equivalent to
translations.
How can nd symmetries of an ODE?. One method is to use the symmetry condition (2.11). In general this is a
complicated nonlinear PDE in ( x, y) . However, Lie symmetries can be derived from a much simpler condition
on the tangent vector eld. By denition, the Lie symmetries of an ODE are of the form
x = x + +O(
2
), y = y + +O(
2
), (2.16)
if we substitute (2.16) into (2.11) we obtain
(x, y) +
_
x
+ (x, y)
y
_
+O(
2
)
1 +
_
x
+ (x, y)
y
_
+O(
2
)
=
_
x + +O(
2
), y + +O(
2
)
_
, (2.17)
17
we now expand each side of (2.17) as a Taylor series about = 0, assuming that each series converges
+
_
x
+ (
y
x
_
2
) +O(
2
) = +
_
x
+
y
_
+O(
2
), (2.18)
where is shorthand for (x, y). Equating the O() terms gives the linearized symmetry condition
x
+ (
y
x
)
y
2
=
x
+
y
. (2.19)
Canonical Coordinates. We know that every ODE whose symmetries include the translations (2.12) may be
integrated directly. How can these coordinates be found?. All orbits of the symmetry (2.12) have the same
tangent vector at every point (, ) = (0, 1) . Given any one-parameter LG, we aim to introduce coordinates
such that (r, s) = (r(x, y), s(x, y)) , where ( r, s) = (r, s + ). If this is possible then, in the new coordinates, the
tangent vector at the point (r, s) is (0, 1) i.e.
d r
d |=0
= 0,
d s
d |=0
= 1, (2.20)
using the chain rule and (2.16) we get
r
x
+ r
y
= 0, s
x
+ s
y
= 1, (2.21)
i.e. Xr = 0, Xs = 1. Since the change of coordinates should be invertible in some neighborhood of (x, y) then we
impose the condition (2.3).
Suppose that we have been able to nd a nontrivial Lie symmetry (LS) of a given ODE (2.8) then this ODE can
be reduce to a quadrature by rewriting it in terms of canonical coordinates as follows:
ds
dr
=
s
x
+ (x, y) s
y
r
x
+ (x, y) r
y
, (2.22)
the right-hand side of (2.22) can be written as a function of r and s. For a general change of variables (x, y)
(r, s) the transformed ODE would be of the form
ds
dr
= (r, s) , (2.23)
for some function . However, (r, s) are canonical coordinates and so the ODE is invariant under the group the
translations in the s-direction: ( r, s) = (r, s + ) . Therefore, as we already know, ODE (2.23) is of the from
ds
dr
= (r) , (2.24)
the problem is now reduced to a quadrature. The general solution of (2.24) is
s
_
(r) dr = c, c R, (2.25)
and hence the general solution of (2.8) is
s(x, y)
_
r(x,y)
(r) dr = c. (2.26)
In general, for an ODE
y
(n)
= (x, y, y
, ....., y
(n1)
), y
(k)
=
d
k
y
dx
k
, (2.27)
where is a smooth function of its arguments. We begin imposing the symmetry condition. A symmetry of
(2.27) is a diffeomorphism. that maps the set of solutions of the ODE to itself. Any diffeomorphism. : (x, y)
( x, y) maps smooth planar curves to smooth planar curves. This action of on the plane induces an action on the
derivatives y
(k)
which is the mapping : (x, y, y
, ....., y
(n1)
) ( x, y, y
, ....., y
(n1)
). This mapping is called
the nth prolongation of . The functions y
(k)
are calculated recursively as follows:
y
(k)
=
d y
(k1)
d x
=
D
x
y
(k1)
D
x
x
, y
(0)
y, (2.28)
where D
x
is the total derivative. The symmetry condition for the ODE (2.27) is
y
(n)
= ( x, y, y
, ....., y
(n1)
), when (2.27) holds. (2.29)
The LS are obtained linearizing (2.29) about = 0. The linearized symmetry condition is derived as above. For
sufciently close to zero, the prolonged LS are of the form:
x = x + +O(
2
), y = y + +O(
2
), y
(k)
= y
(k)
+
(k)
+O(
2
), (2.30)
if we substitute (2.30) into the symmetry condition (2.29) the O() terms yield the linearized symmetry condition
(k)
=
x
+
y
+
(1)
y
+ ..... +
(n1)
y
(n1)
, when (2.27) holds. (2.31)
The functions
(k)
are calculated recursively from (2.28) as follows. For k = 1,
y
(1)
=
D
x
y
D
x
x
=
y
+ D
x
+O(
2
)
1 + D
x
+O(
2
)
= y
+
_
D
x
y
D
x
_
+O(
2
), (2.32)
therefore from (2.30)
(1)
= D
x
y
D
x
. (2.33)
Similarly
y
(k)
=
y
(k)
+ D
x
(k1)
+O(
2
)
1 + D
x
+O(
2
)
, (2.34)
and hence
(k)
= D
x
(k1)
y
(k)
D
x
. (2.35)
To nd the Lie point symmetries of an ODE (2.27) we must rst calculate
(k)
, k = 1, 2, ..., n. The functions and
depend upon x and y only and therefore (2.33) and (2.35) give the following results:
(1)
=
x
+ (
y
x
)y
y
y
2
, (2.36)
(2)
=
xx
+ (2
xy
xx
)y
+ (
yy
2
xy
)y
2
yy
y
3
+ (
y
2
x
3
y
y
)y
, (2.37)
(3)
=
xxx
+ (3
xxy
xxx
)y
+ 3(
xyy
xxy
)y
2
+
_
yyy
yy
_
y
3
yyy
y
4
+ 3(
xy
xx
+
_
yy
3
xy
_
y
2
yy
y
2
)y
3
y
y
2
+
_
y
3
x
4
y
y
_
y
, (2.38)
etc....
For example, for an ODE y
= (x, y, y
) the linearized symmetry condition is obtained substituting (2.36) and

(2.37) into (2.31) and then replacing y
by (x, y, y
). This gives
(2)
=
x
+
y
+
(1)
y
,
i.e.
xx
+ (2
xy
xx
)y
+ (
yy
2
xy
)y
2
yy
y
3
+ (
y
2
x
3
y
y
)
=
x
+
y
+
_
x
+ (
y
x
)y
y
y
2
_
y
. (2.39)
19
For a rst-order ODEs, the right-hand side of the linearized symmetry condition (2.31) is X where X is the
innitesimal generator i.e. X =
x
+
y
. This innitesimal generator is associated with the tangent vector to
the orbit passing through (x, y) , namely
(, ) =
_
d x
d
,
d y
d
_
|=0
. (2.40)
To deal with the action of LS on derivatives of order n or smaller we introduce the prolonged innitesimal generator
X
(n)
=
x
+
y
+
(1)
y
+ ..... +
(n1)
y
(n1)
, (2.41)
the coefcient of
y
(k)
and so X
(n)
is associated with the tangent vector in the space variables (x, y, y
, ....., y
(n1)
).
We can use the prolonged innitesimal generator to write the linearized symmetry condition (2.31) in a compact
form:
X
(n)
(y
(n)
(x, y, y
, ....., y
(n1)
)) = 0, when (2.27) holds. (2.42)
Ejemplo 2.0.3 Consider a third-order ODE
y
= N(x, y, y
, y
). (2.43)
A vector eld S = (x, y)

x
+ (x, y)

y
, is called a symmetry of (2.43) if and only if
S
[3]
(y
N(x, y, y
, y
))|
(2.43)
= 0, (2.44)
where S
[3]
is the third prolongation of S dened by: S
[3]
= S +
3
i=1

(i)
y
(i)
, where
(1)
=
x
+ (
y
x
)y
y
y
2
,
(2)
=
xx
+ (2
xy
xx
)y
+ (
yy
2
xy
)y
2
yy
y
3
+ (
y
2
x
3
y
y
)y
(3)
=
xxx
+ (3
xxy
xxx
)y
+ 3(
xyy
xxy
)y
2
+
_
yyy
yy
_
y
3
yyy
y
4
+ 3(
xy
xx
+
_
yy
3
xy
_
y
2
yy
y
2
)y
3
y
y
2
+
_
y
3
x
4
y
y
_
y
,
and |
(2.43)
means evaluated on (2.43). Straightforward calculations indicate that (2.44) is equivalent to
(3)
= N
x
+ N
y
+
(1)
N
y
+
(2)
N
y
. (2.45)
Since and are independent of y
, Equation (2.45) can be separated with respect to powers of y
to give an overdetermined
system of linear partial differential equations for and . This system is easier to solve than the original equation (2.43):
this is the main stength of the Lie method.
Reduction of order by using canonical coordinates. Suppose that X is an innitesimal generator of a one
paramenter LG of symmetries of the ODE
y
(n)
=
_
x, y
, y
, ...., y
(n1)
_
, n 2. (2.46)
Let (r, s) be canonical coordinates for the group generated by X, so that
X =
s
. (2.47)
If the ODE (2.46) is written in terms of canonical coordinates, it is of the form
s
(n)
=
_
r, s, s, ..., s
(n1)
_
, s
(k)
=
d
k
s
dr
k
, (2.48)
for some , and where s =
ds
dr
. However, the ODE (2.48) is invariant under the LG of translations in s, so the
symmetry condition gives,
s
= 0. Therefore
s
(n)
=
_
r, s, ..., s
(n1)
_
. (2.49)
By writing the ODE (2.46) in terms of the canonical coordinates, we have reduced it to an ODE of order n 1,
for v = s :
v
(n1)
=
_
r, v, ..., v
(n2)
_
, v
(k)
=
d
k+1
s
dr
k+1
. (2.50)
Suppose that the reduced ODE has the general solution
v = f (r; c
1
, c
2
, ...., c
n1
) , (2.51)
then the general solution of the original ODE is
s (x, y) =
_
r(x,y)
f (r; c
1
, c
2
, ...., c
n1
) dr + c
n
. (2.52)
More generally, if v is any function of s and r such that v
s
(r, s) = 0, the ODE (2.49) reduces to an ODE of the
form
v
(n1)
=

_
r, v, ..., v
(n2)
_
, v
(k)
=
d
k
v
dr
k
. (2.53)
Once the general solution of (2.53) has been found, the relationship s = s (r, v) , gives the general solution of
(2.46):
s (x, y) =
_
r(x,y)
s (r; v (r; c
1
, c
2
, ...., c
n1
)) dr + c
n
. (2.54)
Ejemplo 2.0.4 Consider the linear ODE
y
=
_
3
x
2x
_
y
+ 4y (2.55)
where by using all the above theoretical results we get
yy
= 0,
2
_
3
x
2x
_
y
+
yy
2
xy
= 0,
12y
y
xx
+
_
3
x
2
+ 2x
_

_
3
x
2x
_
x
+ 2
xy
= 0,
8y
x
4 + 4y
y
_
3
x
2x
_
x
+
xx
= 0,
and therefore we obtaing the following symmetries:
X
1
= y
y
, X
2
=
_
x
2
1
_
y
, X
3
= e
x
2
y
, X
4
=
y
x
3
2y
2
x
2

y
,
X
5
=
x
2
1
x
3

x
+
2y
x
2
y
, X
6
=
e
x
2
x
3

x
2ye
x
2
x
2

y
,
X
7
=
y(x
2
1)e
x
2
x
3

x
2y
2
e
x
2
x
2

y
, X
8
=
(x 1)
2
(x + 1)
2
e
x
2
x
3

x
2y (x 1) (x + 1) e
x
2
x
2

y
.
21
We consider the the rst of them, namely X
1
= y
y
, where the simplest canonical coordinates are
r = x, s = ln y,
which prolong to
v =
ds
dr
=
y
y
,
d
2
s
dr
2
=
y
y

_
y
y
_
2
,
hence our ODE reduces to the Riccati equation
v
=
_
3
r
2r
_
v + 4 v
2
,
where its solution is
v =
2r
_
C
1
e
r
2
1
_
C
1
e
r
2
1 + r
2
,
all that remains it to carry out the c.v.
y
y
=
2x
_
C
1
e
x
2
1
_
C
1
e
x
2
1 + x
2
,
so the solution for our original ODE yields
y = C
2
_
C
1
e
x
2
1 + x
2
_
.
Ejemplo 2.0.5 For the ODE
y
=
y
2
y
+
_
y
1
y
_
y
we nd the following symmetry

X =
x
which leads us to get the following c.v.
r = y, s(r) = x,
thus (this is a central step in all the work)
v =
ds
dr
=
1
y
, v
=
y
(y
)
3
,
and therefore we get:
v
=
v
r

_
r +
1
r
_
v
2
.
Bibliography
[1] G. W. Bluman and J. D. Cole. Similarity Methods for Differentials Equations. (1974) Appl. Math. Sci. N13
Springer-Verlang New York. G. W. Blumann and S.W. Kumei Symmetries and Differential Equations.
Springer Verlang (1989). G. W. Blumann and S.C. Anco. Symmetries and Integration Methods for Differen-
tial Equations". Springer Verlang (2002).
[2] L. V. Ovsiannikov, Group Analysis of Differential Equations (Academic Press 1982).
[3] N. H. Ibragimov. Transformation Groups Applied to Mathematical Physics. D. Reidel (1985). N. H. Ibragimov.
Elementary Lie Group Analysis and Ordinary Differential Equations (John Wiley & Sons, 1999).
[4] R. Seshadri and T. Y. Na. Group Invariance in Engineering Boundary Value Problems. Springer-Verlang. NY,
1985.
[5] H. Stephani, Differential Equations: Their Solutions Using Symmetries (Cambridge University Press 1989).
[6] P. T. Olver, Applications of Lie Groups to Differential Equations (Springer-Verlang, 1993).
[7] L. Dresner, Applications of Lies Theory of Ordinry and Partial Differential Equations. IOP 1999
[8] G. Bauman, Symmetry Analysis of Differential Equations. Springer Verlang (2000).
[9] P. E. Hydon, Symmetry Methods for Differential Equations (Cambridge University Press, 2000).
[10] B. J. Cantwell, Introduction to Symmetry Analysis (Cambridge University Press, Cambridge, 2002).
23
24 BIBLIOGRAPHY
Chapter 3
Ecuaciones de Primer Orden
3.1 Introduccin
D.A. has usually been employed in different areas (elds) such as engineering problems, uid mechanics etc...
and these problems are always described by partial differential equations (pde) (see [1]-[10]). This method (tac-
tic") helps us to reduce the number of quantities that appear into an equation and to obtain ordinary differential
equations (ode). We would like to point out (emphasize) that this tool is more effective if one practices the spatial
discrimination, such tactic allows us to obtain better results than with the standard application of D.A. (see [11]).
Knowing that D.A. works well in pde we would like to extend this method to the study of ode (the rst order
in this case) in a systematic way. There are in the literature previous work in this direction for ode of rst order
(see [12]-[13].
The method of the Lie groups it has showed as a very useful tool in order to solve nonlinear equations (nl-ode)
as well as pde (see [14]-[27] ). Nevertheless, when one is studying ode of rst order its application results very
complicated (very tedious) if one has not a computer algebra package since if one decides to look for the possible
symmetries of an ode with pencil and paper this task may be turned very exhausting. However, if we know
that an ode admits a concrete symmetry, then it is a trivial issue to nd new variables which allows us to rewrite
the ode in quadratures or as we will see in this paper to obtain an ode with separating variables.
Our purpose in this work is to explain, through examples, how D.A. works in order to nd these changes of
variables (c.v.) in a trivial way, i.e. without the knowledge of the symmetries of the ode under study. The idea
is as follows. When we are studying an ode form the dimensional point of view, we must require that such
ode veries the principle of dimensional homogeneity (pdh) i.e. that each term within the equation have the
same dimensions, for example, speed, or energy density. To clarify this concept we consider the following ode,
y
=
y
x
+ x, where each term must have dimensions of y
i.e. [y
] = yx
1
, where [] stands for dimensional
equation of the quantity . As we can see, this ode does not verify the pdh, since the term x has dimensions of x,
i.e. [x] = x. In order to do that this equation veries the pdh we need to introduce dimensional constants, in such
a way that after rewrite the ode with these constants the ode now veries such principle i.e. in this example and
as we can see easily if we consider only one constant a, such that [a] = yx
2
, we make that the ode y
=
y
x
+ ax,
veries the pdh. We would like to emphasize that this situation does not appear (arise) when one is studying
physical or engineering problems since (as it is supposed) that such problems (equations) verify the pdh and
we do not need to introduce new dimensional constants, that must have physical meaning, for example, the
viscosity coefcient, etc....
Precisely this dimensional constant suggests us the c.v. (x, y(x)) (t, u(t)) where
_
t = x, u = yx
2
_
in such
a way that rewriting the original ode in these new variables it is obtained a new ode with separating variables:
u + u
t = 1. The reason is the following. We know from the Lie group theory that if it is known a symmetry of
an ode then this symmetry bring us through a c.v. to obtain a simpler ode (a quadrature or a ode with separating
25
26 CHAPTER 3. ECUACIONES DE PRIMER ORDEN
variables) and therefore the solution is found in a closed form. To nd this c.v. it is used the invariants that
generate each symmetry, i.e. the rst principle is that it is useful to pass to new coordinates such that one of the
coordinate functions is an invariant of the group. After such transformation it often (but not always) happens
that the variables separate and the equation can be solved in closed form. We must stress that taking the new
dependent variables to be an invariant of the group does not guarantee the separation of variables. The choice
of the independent variable is also very important.
One must note that D.A. gives us this invariant (or at least a particular solution). It is observed in our example
that the solution y = x
2
, (suggested by D.A.) is a particular solution for this ode, but if we study this ode form
the Lie theory point of view it is obtained that such ode has the scaling symmetry X = x
x
+2y
y
and therefore
the solution y = x
2
, is furthermore an invariant solution (generated by X), for this reason such c.v. works well.
As we will show in the paper, the c.v. that D.A. induces is the same than the generated one by the symmetries of
the ode. In order to make see this fact, we will solve each example by D.A. as well as by the Lie group technique
(abusing of the trivial calculations), calculating the symmetries and their corresponding invariants and c.v..
However D.A. has a little (or rather big) drawback, it only gives us relationship of type power i.e. y = x
n
, with
n R, and we cannot obtain relationship as y = (x + 1)
n
. Nevertheless, and as we will see in the examples,
although D.A. does not give us an invariant solution it will be sufcient that it provides us a particular solution
in order to obtain a c.v. which brings us to a simpler ode than the original one. This task will be very useful in
the case of studying Riccati and Abel odes. Furthermore, this tactic" is valid even in the case in which the ode
has no symmetries but unfortunately one always founds examples in which any tactic does not work.
The paper is organized as follows: In section 2 we describe (with all the tedious and superuous calculations),
through two examples, how the proposed method works by comparing it with the Lie method in order to show
which are the symmetries and their corresponding invariants. In section 3 we shall show several kind of ODEs
paying special emphasis in for example Bernoulli, Riccati and Abel ODEs (see [28]-[32]). We think that these
kind of ode are the most difcult to solve and therefore any simplication is welcome. We show two Abel odes
that do not admit symmetries (see [33]-[37]). In the rst of them we use the D.A. to obtain a particular solution
that allows us to obtain a Bernoulli ode but in the second one we are not able to nd any particular solution and
therefore a c.v. which brings us to obtain a simpler ode. We end with some conclusion as well as pointing out
some of the limitations of the proposed tactic.
We would like to emphasize the pedagogical character of this chapter (at least we try) for this reason we have
abused of superuous calculations and we have omitted some technical details since the supposed audience are
mainly engineers and/or physicists but not for mathematicians.
3.2 The method
In this section we will explain how the D.A. works in order to solve odes through two simple examples. The
main idea is to introduce dimensional constants that make the ode under study verify the principle of dimen-
sional homogeneity. These dimensional constants help us to nd c.v. which bring us to obtain ode with sepa-
rating variables. These c.v. obtained through D.A. correspond to invariant solution (or at least to a particular
solution) and therefore they are induced by one symmetry. We compare our tactic with the Lie one.
Ejemplo 3.2.1 Solve the homogeneous equation
_
x
2
+ y
2
_
dx = 2xydy. (3.1)
Solution. In order to solve it we will use three different ways, the traditional, the dimensional and the Lie
method.
3.2. THE METHOD 27
Traditional method. Making the c.v. u = y/x we have:
u
=
1 u
2
2ux
=
2udu
1 u
2
=
dx
x
= ln(1 u
2
) = ln x = y
2
= x
2
+ x, (3.2)
Dimensional Analysis. We go next to consider eq. (3.1), written as follows
y
=
_
a
2
x
2
+ y
2
_
2xy
, (3.3)
where the dimensional constant a, makes the ode verify the dimensional principle of homogeneity (d.p.h.) if
[a] =
_
y
x
_
= X
1
Y, (3.4)
where [] stands for the dimensional equation of the quantity .
Applying the Pi theorem we obtain the dimensionless variables that help us to simplify the original ode. There-
fore taking into account the following dimensional matrix we take
y x a
x 0 1 1
y 1 0 1
=
1
=
ax
y
, = y = ax, (3.5)
as we will see later this solution is a particular solution of this ode. We would like to emphasize that as we have
only needed one constant then the equation is scale-invariant in such a way that the generator of this group is
X = x
x
+ y
y
as it is observed from eq. (3.5). This fact will be probed studying this equation under the Lie
group tactic, see below.
In this way the new variables are (t, u(t)) :
_
t = x, u(t) = a
x
y
_
, =
_
x = t, y = a
t
u(t)
_
, (3.6)
this change of variables brings us to rewrite eq. (3.3) as follows:
u
u (1 u
2
)
=
1
2t
=
u
1 u
2
=
1
2
ln t, (3.7)
and hence
y
2
= a
2
x
2
+ C
1
x. (3.8)
As we can see in this trivial example, the D.A. induces a c.v. which helps us to obtain an ode simpler than the
original one.
We can also think in the following way
ay
= b
x
2y
+ a
y
2x
, (3.9)
where
[a] = x, [b] = y
2
x
1
, (3.10)
and hence
_
t =
x
a
, u(t) =
y
2
bx
_
=
_
x = at, y =
_
abtu(t)
_
, (3.11)
therefore eq. (3.9) yields
u
= 1 = u = t + C
1
, (3.12)
in this way we obtain the solution
y
2
=
b
a
x
2
+ C
1
x, (3.13)
once we have obtained the solution, the constants a, b are makig equal to 1, i.e. a = b = 1.
In the same way we may consider the following change of variables
_
t =
x
a
, u(t) =
y
bx
_
=
_
x = at, y = u(t)
abt
_
, (3.14)
which brings us to the following ode
2u
u = 1 = u
2
t + C
1
= 0, (3.15)
and therefore we obtain again the solution (3.13). As we will see below all these c.v. are generated by their
corresponding (respective) symmetry.
As we can see, this last change of variables is better than the rst and the tactic is the same: to introduce dimen-
sional constants that make the equation dimensional homogeneous.
Lie Method. In order to nd the symmetry generator of a ode
y
= f (x, y) , (3.16)
we need to solve the following pde
x
+ (
y
x
) f
y
f
2
(x, y) f
x
(x, y) f
y
= 0, (3.17)
where
f
x
=
d f
dx
, f
y
=
d f
dy
. (3.18)
In this case we have to solve
x
+
1
2
(
y
x
)(x
2
+ y
2
)
x y

1
4
y
(x
2
+ y
2
)
2
x
2
y
2
(x, y)(
1
y

x
2
+ y
2
2 x
2
y
) (x, y)(
1
x

x
2
+ y
2
2 x y
2
) = 0, (3.19)
we nd that
_
X =
x
+
y
_
:
X
1
=
x
2 y
y
, X
2
=
_
x
2
2 y
+
y
2
_
y
, X
3
= x
x
+ y
y
,
X
4
=
x
+
y
2 x
y
, X
5
=
_
x
2
+ y
2
_
x
+ 2xy
y
, (3.20)
observing that X
5
is a trivial symmetry.
Each symmetry induces a change of variable (canonical variables) which are obtained through the following
formula
Xt = 0, Xu = 1. (3.21)
In this way the change of variables that induces the eld X
4
=
x
+
y
2 x
y
is:
_
t =
y
x
, u(t) = x
_
=
_
x = u(t), y = t
_
u(t)
_
, (3.22)
in such a way that eq. (3.75) yields:
u
= 2t = u(t) = t
2
+ C
1
, (3.23)
and therefore
x =
y
2
x
+ C
1
, (3.24)
which is the same solution. The c.v. induced by this symmetry is similar to the obtained one in (3.14).
3.2. THE METHOD 29
The c.v. that induces the eld X
1
=
x
2 y
y
is the following one
_
t = x, u(t) =
y
2
x
_
=
_
x = t, y =
_
u(t)t
_
, (3.25)
therefore eq.(3.1) yields
u
= b
2
, (3.26)
this c.v. is the same than the obtained one in (3.11).
For example the c.v. that induces the eld X
3
= x
x
+ y
y
is
_
t =
y
x
, u(t) = ln x
_
=
_
x = e
u(t)
, y = te
u(t)
_
, (3.27)
which brings us to the following ode :
u
=
2t
t
2
1
= u(t) = ln (1 + t) ln (1 + t) + C
1
, (3.28)
now writing the solution in the original variables we take
ln x = C
1
+ ln
_
x
(y x) (y + x)
_
. (3.29)
Now if we consider the invariants that induce each symmetry
dx
=
dy
, y
:=
dy
dx
=

, (3.30)
we take that:
X
1
I
1
= x, X
2
I
2
= x, X
3
I
3
=
y
x
, X
4
I
4
=
y
x
. (3.31)
For example the symmetry X
3
= x
x
+ y
y
generates the following invariant:
dx
=
dy
=
dx
x
=
dy
y
= ln x = ln y = I
3
=
y
x
y = ax, (3.32)
this would be the solution that suggests us precisely the direct use of the Pi theorem (if the ode is scale invariant,
as in this case, the solution obtained applying by the Pi theorem coincides with the invariant solution that
generates the scale symmetry, in this case X
3
). We see that we only obtain a particular solution, but that this is
invariant, in fact if we think about the ode as a dynamical system we see that the xed point of such equation
would be precisely the solution y = x.
Ejemplo 3.2.2 Solve the linear ode
_
1 x
2
_
y
+ xy = 1. (3.33)
Solution. Solution through D.A. Our rst step will be to introduce dimensional constants in such a way that
eq. (3.33) veries the principle of dimensional homogeneity. In this way we rewrite the equation as follows
_
A x
2
_
y
+ xy = B, (3.34)
where
[A] = X
2
, [B] = XY, (3.35)
i.e.
1
=
x
2
A

2
=
xy
B
, (3.36)
where
y =
1
Bx
_
x
2
A
_
, (3.37)
being a unknown function. It is possible to make the following assumption
y =
1
Bx
_
x
2
A
_
n
, n R. (3.38)
Since y = x
1
is not a particular solution of (3.33) we need to combine the monomias in order to nd a
particular solution, the simplest way is as follows:
_
t =
x
2
A
, u(t) =
A
B
y
x
_
=
_
x =
At, y =
B
A
u(t)
At
_
(3.39)
observing that
u(t) =
2

1
1
, (3.40)
where y = cx, derived from
2

1
1
is a particular solution of (3.33).
Observacin 3.2.1 (Recipe) When we have two monomia, we have to check if they induce any particular (or invari-
ant) solution. If they are not particular solutions then we must combine them in order to obtain such solution, this solution
usually is obtained by combining them in a very simple way.
In this way our ode is written now as follows:
2u
u 1
=
1
t (t 1)
= u = 1 + C
1
t + 1
t
, (3.41)
and in the original variables the solution of eq. (3.33) yields:
y =
A
B
_
x C
1
_
A x
2
_
, (3.42)
where C
1
is an integration constant. Now making A = B = 1, we have the ordinary solution to eq. (3.33).
Lie method. Following the standard procedure we have to solve the pde:
x
+
(
y
x
) (1 x y)
1 x
2

y
(1 x y)
2
(1 x
2
)
2
(x, y) (
y
1 x
2
+
2 (1 x y) x
(1 x
2
)
2
) +
(x, y) x
1 x
2
= 0, (3.43)
which solutions are:
X
1
=
_
1 + x
2
y
, X
2
= (y + x)
y
, X
3
=
_
(2xy y
2
1)(y + x)
(x 1)(x + 1)
_
y
, (3.44)
and their corresponding invariants are:
X
1
I
1
= x, X
2
I
2
= x, X
3
I
3
= x. (3.45)
For example the symmetry X
1
generates the following c.v.:
_
t = x, u(t) =
y
1 + x
2
_
=
_
x = t, y = u(t)
_
1 + t
2
_
, (3.46)
3.3. CUADRATURAS. 31
therefore eq. (3.33) is written as:
u
1 + t
2
1 2t
2
+ t
4
, (3.47)
which solution is:
u(t) =
_
1 + t
2
_
3/2
4 (t + 1)
2
+
_
1 + t
2
_
3/2
4 (t 1)
2
+ C
1
, (3.48)
and hence
y
x
2
1
=
C
1
(x
2
1) + x
x
2
1
x
2
1
= y = x c
x 1
x + 1. (3.49)
Now, if for example we consider the symmetry X
2
then we have:
(t = x, u(t) = ln(y + x)) =
_
x = t, y =
te
u(t)
1
e
u(t)
_
, (3.50)
therefore:
u
=
t
1 + t
2
, (3.51)
nding in this way that its solution is:
u(t) =
1
2
ln(t 1)
1
2
ln(t + 1) + C
1
, (3.52)
hence in the original variables the solution yields:
ln(y + x) =
1
2
ln(x 1)
1
2
ln(x + 1) + C
1
, (3.53)
which after a simple simplication it yields:
y = x c
x 1
x + 1, (3.54)
as we already know.
With these two simple examples we have tried to show how the D.A. works in order to introduce c.v. which
brings us to obtain simpler odes than the original one. As we have emphasized in the recipe, the trick is to look
for a particular solution. Sometimes this particular solution will be furthermore invariant solution (induced by
a concrete symmetry). If this is the case, then our ode will be reduced to an ode with separating variables but if
this solution is only a particular then, as we will see below, we have not any guarantee of reducing our ode to an
ode with separate variables, nevertheless we will obtain a simpler ode than the original one. In the next section
we will study some examples.
3.3 Cuadraturas.
Decimos que una ecuacin diferencial est en forma de cuadraturas cuando sucede alguna de las siguientes
situaciones:
1. la ecuacin es de primer orden y el lado derecho de la ecuacin depende slo de x
y
= F(x) (3.55)
la solucin es por lo tanto de la forma
y(x) =
_
F(x) dx + C
1
. (3.56)
2. la ecuacin es de primer orden y el lado derecho de la ecuacin depende slo de y(x):
y
= F(y(x)) (3.57)
la solucin es por lo tanto de la forma
_
y(x)
1
F(a)
da + C
1
= x (3.58)
3.3.1 Simetras.
Desde el punto de vista de las simetras vemos que la ecuacin (3.55) slo puede tener las siguientes simetras
x
+ (
y
x
) F(x)
y
F(x)
2
(x, y) F
= 0
encontrando que:
[ = 0, = 1]
mientras que la ecuacin (3.57) slo puede tener las simetras.
x
+ (
y
x
) F(y)
y
F(y)
2
(x, y) F
y
= 0
encontrando que:
[ = 1, = 0]
3.3.2 Ejemplos
Veremos unos sencillos e inmediatos ejemplo de cada tipo de ecuaci descrito en la anterior subseccin.
Ejemplo 3.3.1 Resolver la ecuacin en cuadraturas
x + x y
2
= 1
Solucin. Vemos que
y =
_
x
2
+ x
1
2
arcsin(2 x 1) + C
1
,
y =
_
x
2
+ x +
1
2
arcsin(2 x 1) + C
1
ya que
x
+
(
y
x
)
_
x (x 1)
x
+
y
(x 1)
x
(x, y) (
_
x (x 1)
x
2
+
2 x + 1
2 x
_
x (x 1)
) = 0
y por lo tanto
[ = 0, = 1]
como ya sabamos.
Intentaremos obtener alguna solucin particular a la ecuacin
y
=
_
a bx
x
3.4. ECUACIONES EN VARIABLES SEPARABLES Y REDUCIBLES A ELLAS. 33
donde las constantes dimensionales a y b tienen las siguientes dimensiones: [a] = y
2
x
1
, [b] = y
2
x
2
. En este
caso la solucin que sugiere el AD es de la forma
y
2
= ax
_
bx
2
y
2
_
,
observando que semejante sugerencia est muy lejos de ser una solucin a la ODE propuesta. Intentaremos otra
tctica dimensional, forzando a que la funcin desconocida sea de la forma
n
b
, i.e.
y
2
= ax
_
bx
2
y
2
_
n
= y
2n+2
= ab
n
x
2n+1
de esta forma
y =
_
ab
n
x
2n+1
_ 1
2n+2
que tampoco funciona.
Ejemplo 3.3.2 Resolver la ecuacin en cuadraturas
y
= y
2
Solucin. Vemos que
y =
1
x C
1
ya que
x
+ (
y
x
) y
2
y
y
4
2 (x, y) y = 0
y por lo tanto
[ = 1, = 0]
como ya sabamos.
En este ejemplo la solucin que nos sugiere el AD es la siguiente. Introducimos la constante a para hacer que la
EDO sea d.h. i.e. y
= ay
2
, con [a] = y
1
x
1
, de esta forma llegamos a
y a x
y 1 1 0
x 0 1 1
= y =
1
ax
observando que slo despus de haber sustituido esta solucin en la ODE original nos daramos cuenta del signo
i.e. y =
1
ax
, ya que el AD no entiende ni de signos ni de constantes numricas, slo de rdenes de magnitud.
Claim 3.3.1 Resaltamos que a pesar de no tener simetras de escala hemos podido obtener una solucin por AD.
3.4 Ecuaciones en variables separables y reducibles a ellas.
Son de la forma
y
= f (x)g(y) (3.59)
3.4.1 Mtodo tradicional.
Tenemos que separar, viendo de forma inmediata que:
y
g(y)
= f (x) (3.60)
por lo tanto
dy
g(y)
= f (x)dx =
_
dy
g(y)
=
_
f (x)dx + C
1
(3.61)
3.4.2 Mtodo dimensional.
En primer lugar deberemos reescribir la ecuacin introduciendo cuantas constantes dimensionales sena nece-
sariaas para hacer que la ecuacin verique el principio de homogeneidad de tal forma que
_
dy
g(y)
_
= [A
1
f (x)dx] (3.62)
donde la constante A
1
hace que la anterior ecuacin sea dimensionalmente homognea.
3.4.3 Mtodo de Lie.
En esta ocasin el mtodo de Lie nos lleva a obtener las siguientes simetras para la ecuacin (3.59) al resolver la
siguiente ecuacin:
x
+ (
y
x
) f (x)g(y)
y
f (x)
2
g(y)
2
(x, y) f
g(y) (x, y) f (x)g
= 0 (3.63)
cuyas soluciones son:
[ = 0, = g(y)], [ =
1
f (x)
, = 0], [ =
1
f (x)
, = g(y)]
que conducen a la siguiente solucin:
dx
(x, y)
=
dy
(x, y)
=
_
dy
g(y)
=
_
f (x)dx + C
1
Resaltamos que de las simetras obtenidas =
1
f (x)
, = g(y), podremos tener simetras de escala (las que genera
el AD) sii f (x) =
1
x
y g(y) = y.
3.4.4 Ejemplo.
Ejemplo 3.4.1 Vamos a estudiar la siguente ode:
y
= sin(x)y
Solucin. Resolveremos la ecuacin planteada por los tres mtodos, intentando mostar las ventajas y desventa-
jas de cada mtodo.
3.4. ECUACIONES EN VARIABLES SEPARABLES Y REDUCIBLES A ELLAS. 35
1. Mtodo tradicional.
Vemos de forma inmediata que la solucin es:
_
dy
y
=
_
sin xdx = y = C
1
e
cos(x)
2. Mtodo dimensional.
Tenemos que introdir constantes para hacer que la ecuacin sea dimensionalmente homognea, encon-
trando as que:
y
= A
1
sin(x)y
donde [A
1
] = x
1
y donde estamos considerendo que [sin(A
1
x)] = x
0
y
0
, i.e. es adimensional. Por lo tanto
la soluin que sugiere el AD es:
y = y
0
(sin(A
1
x))
donde hace referencia a una uncin desconocida de sin(A
1
x). Con este ejemplo vemos que en este caso
el AD no nos acerceca ni de lejos a una solucin a una ecuacin tan sencilla como la planteada.
3. Mtodo de Lie.
Veamos ahora las simetras de la ecuacin al resolver la ecuacin:
x
(
y
x
)y sin(x)
y
y
2
sin(x)
2
+ (x, y) y cos(x) + (x, y) sin(x) = 0
todas las posibles simetras son:
[ = 0, = y], [ = 0, = e
cos(x)
], [ =
1
sin(x)
, = 0], [ =
1
sin(x)
, = y].
por lo tanto:
_
sin xdx =
_
dy
y
llegando as a la solucin:
y = C
1
e
cos(x)
Las variables cannicas generadas por [ = 0, = y] nos llevan al siguiente resultado:
(t = x, u(t) = ln y)
_
x = t, y = e
u(t)
_
por lo tanto la ecuacin diferencial original queda reducida a:
u
= sin t = u = cos t + C
1
por lo tanto la solucin en las variables originales es:
ln y = cos x + C
1
= y = C
1
e
cos(x)
.
Podemos observar las ventajas y desventajas que presentan cada mtodo.
Ejemplo 3.4.2 Vamos a estudiar la siguente ode:
y
= xy
Solucin. Resolveremos la ecuacin planteada por los tres mtodos, intentando mostar las ventajas y desventa-
jas de cada mtodo.
Vemos de forma inmediata que la solucin es:
_
dy
y
=
_
xdx = y (x) = C
1
e
1
2
x
2
2. Mtodo dimensional.
Como en l ejemplo anterior reescribimos la ecuacin de tal forma que
y
= A
1
xy
con [A
1
] = x
2
y = y
0
_
A
1
x
2
_
bastante pobre.
3. Mtodo de Lie.
Veamos ahora las simetras de la ecuacin al resolver la ecuacin:
x
(
y
x
)yx
y
y
2
x
2
+ (x, y)y + (x, y)x = 0
una de las posibles simetras es:
[ = 0, = y], [ = 0, = e
1
2
x
2
], [ =
1
x
, = 0], [ =
1
x
, = y].
por lo tanto:
_
xdx =
_
dy
y
llegando as a la solucin:
y (x) = C
1
e
1
2
x
2
El empleo de las variables cannicas (v.c) que generan la simetra [ = 0, = y] nos lleva a los siguientes
resultados:
(t = x, u(t) = ln y)
_
x = t, y = e
u(t)
_
por lo tanto la ecuacin diferencial original queda reducida a:
u
= t = u =
1
2
t
2
+ C
1
por lo tanto la solucin en las variables originales es:
ln y =
1
2
x
2
+ C
1
= y (x) = C
1
e
1
2
x
2
.
3.5 Ecuacin homognea
En primer lugar deberemos distinguir los siguientes casos:
1. Homognea clase A
dy
dx
= f
_
y
x
_
(3.64)
3.5. ECUACIN HOMOGNEA 37
2. Homognea clase B
F
_
y
,
y
x
_
(3.65)
3. Homognea clase C
y
= F
_
ax + by + c
rx + sy + t
_
(3.66)
4. Homognea clase D
y
=
y
x
+ g(x) f
_
y
x
_
(3.67)
5. Homognea clase G
y
=
y f
_
y
x
a
_
x
(3.68)
Describos a continuacin los mtodos de resolucin de cada tipo de ecuacin.
1. Clase A. La forma de resolver la ecuacin (3.64) es la de hacer el tpico c.v.
u =
y
x
(3.69)
de tal forma que la ecuacin a resolver es:
xu
= f (u) u =
du
f (u) u
=
dx
x
(3.70)
que es de variables separables y por lo tnato facilmente integrable.
2. Clase B. La forma de resolver la ecuacin (3.65) es la de hacer el tpico c.v.
(t) =
y
x
, y
= (t) (3.71)
de esta forma derivando y = x(t) respecto de x aparece una ecuacin en variables separadas.
3. Clase C. La forma de resolver la ecuacin (3.66) deberemos tener en cuenta las siguientes posibilidades:
(a) Que las dos rectas ax + by + c y rx + sy + t se corten en un punto (x
0
, y
0
) . En este caso se efectua el
c.v. X = x x
0
, Y = y y
0
. De esta forma la ecuacin resultante se reduce a una homognea de clase
A.
(b) Que las rectas ax + by + c y rx + sy + t sean paralelas. En este caso se efectua el c.v. z = ax by. De
esta forma la ecuacin resultante es de variables separadas.
4. Clase D. La forma de resolver la ecuacin (3.67) es la de hacer el tpico c.v.
5. Clase G. La forma de resolver la ecuacin (3.68) es la de hacer el tpico c.v.
La ecuacin (3.64) ser homognea dimensionalmente si no se discrimina entre x e y, en cuyo caso
=
y
A
1
x
= (3.72)
es adimensional, [A
1
] = yx
1
.
Expresndola en funcin de y de x, tendremos y = A
1
x,
y
= + A
1
x
d
dx
, (3.73)
eliminando y
, y en (3.64) obtenemos
+ A
1
x
d
dx
= f () =
dx
x
=
d
f () 1
, (3.74)
consiguindose as la separacin de variables que permite su integracin. Como se observa es el msmo mtodo
que el clsico.
3.5.3 Mtodo de Lie.
Como en el caso del mtodo tradicional distinguiremos cada tipo:
1. Clase A. Tenemos que resolver la ecuacin
x
+ (
y
x
) f (
y
x
)
y
f (
y
x
)
2
+
(x, y)D( f )(
y
x
)y
x
2

(x, y)D( f )(
y
x
)
x
= 0
que nos conduce a:
[ = 0, = y + x f (
y
x
)], [ = x, = y]
y por lo tanto la osolucin general ela ecuacin es:
_
y/x
1
a f (a)
da + ln(x) + C
1
= 0
Resaltamos que esta tipo de ecuaciones admite simetras de escala [ = x, = y] (AD) y que por lo tanto la
menos seremos capaces de encontrar soluciones particulares (o la general comleta en el mejor de los casos)
por medio de AD.
2. Clase B. En este caso las simetras que genera la ecuacin son:
[ = x, = y]
y la solucin general es del tipo:
_
y/x
1
F(
y
x
, a) + a
da + ln(x) + C
1
= 0
3. Clase C. La ecuacin a resolver es:
x
+ (
y
x
)F
_
ax + by + c
rx + sy + t
_
y
F
_
ax + by + c
rx + sy + t
_
2
+
+(x, y)D( f )
_
ax + by + c
rx + sy + t
_
_
a
rx + sy + t

(ax + by + c) r
(rx + sy + t)
2
_
(x, y)D( f )
_
ax + by + c
rx + sy + t
_
_
b
rx + sy + t

(ax + by + c) s
(rx + sy + t)
2
_
= 0
encontrando que las simetras en este caso son:
[ =
x(br sa) + bo sc
br sa
, =
rc ao + (rb as)y
br sa
]
que conducen a la siguiente solucin (invariantes):
dx
(x, y)
=
dy
(x, y)
=
dx
x(br sa) + bo sc
=
dy
rc ao + (rb as)y
,
4. Clase D. El sistema a resolver es:
x
+ (
y
x
)
_
y
x
+ g(x) f
_
y
x
__
y
_
y
x
+ g(x) f
_
y
x
__
2
(x, y)
_
y
x
2
+ g
(x) f
_
y
x
_
gD( f )
_
y
x
_
y
x
2
_
(x, y)
_
1
x
+
gD( f )
_
y
x
_
x
_
= 0
encontrando que las simetras en este caso son:
[ = 0, = x f (
y
x
)], [ =
x
g(x)
, =
y
g(x)
]
que conducen a la siguiente solucin (invariantes):
dx
(x, y)
=
dy
(x, y)
=
g(x)dx
x
=
g(x)dy
y
,
sin embargo, si efectuamos el siguiente c.v.
cv := {x = t, y = u(t) t}
encontramos que
u
t
t + u(t) = u(t) + g(t) f (u(t))
cuya solucin es
_
g(t)
t
dt
_
u(t)
1
f (a)
da + C
1
= 0
y por lo tanto
_
y/x
1
f (a)
da +
_
g(x)
x
dx + C
1
= 0
5. Clase G. El sistema a resolver es:
x
+ (
y
x
)
_
y f
_
y
x
a
_
x
_
y
_
y f
_
y
x
a
_
x
_
2
(x, y)
_
y
2
Df
_
y
x
a
_
a
x
a
x
2

f
_
y
x
a
_
y
x
2
_
(x, y)
_
Df
_
y
x
a
_
x

yDf
_
y
x
a
_
x
a
x
2
_
= 0.
Las simetras que encontramos estn generadas por:
[ = 0, = y (f (
y
x
n
) + n)], [ = x, = y n],
si efectuamos el siguiente c.v.
cv := {x = t, y = u(t) t},
encontramos una nueva ecuacin diferencial
u
t
t + u(t) = u(t) f (
u(t) t
t
n
),
cuya solucin es:
ln(t) C
1
+
_
u(t)/n1
1
a(n f (a))
da,
volviendo a las variables originale sencontramos que:
ln(x) + C
1
+
_
y/x
n
1
a( f (a) n)
da = 0.
3.5.4 Ejemplos.
Ejemplo 3.5.1 Queremos integrar la ecuacin homognea
_
x
2
+ y
2
_
dx = 2xydy (3.75)
Solucin. Utilizaremos tres tcticas.
Haciendo el c.v. u = y/x entonces
u
=
1 u
2
2ux
=
2udu
1 u
2
=
dx
x
= ln(1 u
2
) = ln x = y
2
= x
2
+ x
2. Solucin mediante Anlisis dimensional.
Consideramos, en vez de (3.75), la ecuacin
_
b
2
x
2
+ y
2
_
dx = 2xydy (3.76)
que ser discriminadamente homognea (d.h.) si
[b] =
_
y
x
_
(3.77)
Obtendremos como variable que reemplaza a y la adimensional , obtenida obviamente de la siguiente
matriz (ec. 3.78 )
b x y
L
x
1 1 0
L
y
1 0 1
(3.78)
De manera que podemos escribir
= b
x
y
, y = b
x
,
dy
dx
= b
x
2
d
dx
+
b
=
b
_
1
x
d
dx
_
. (3.79)
Sustituyendo y,
dy
dx
en (3.76)
b
2
_
x
2
+
x
2
2
_
=
2bx
2
_
1
x
d
dx
_
(3.80)
de donde resulta
2
+ 1 =
2x
d
dx
. (3.81)
Separando variables
dx
2x
=
d
_
1
2
_ =
d
+
d
1
2
= (3.82)
1
2
ln x + ln A = ln
1
2
ln
_
1
2
_
= A
x =

_
1
2
, (3.83)
y sustituyendo = b
x
y
resulta tras sencillas operaciones
y
2
= b
2
_
x
2
+
x
A
2
_
(3.84)
que es la solucin de (3.76). La solucin de (3.75) corresponde a b = 1 y ser
y
2
= x
2
+ cx (3.85)
siendo c una constante arbitraria.
Tambin podemos seguir este otro mtodo dimensional alternativo al anterior. Reescribimos la ecuacin
(3.75) como
y
=
_
c
1
x
2
+ y
2
_
2xy
, (3.86)
donde c
1
es cierta constante con dimensiones [c
1
] = y
2
x
2
. De esta forma intentamos encontar una solucin
aplicando directamente el teorema Pi a las magnitudes (y, x, c
1
) , obteniendo:
y x c
1
x 0 1 2
y 1 0 2
= y = c
1/2
1
x, (3.87)
viendo de esta forma que hemos obtenido una solucin particular a nuestra ecuacin diferencial por un
mtodo pedestre, comprobamos que se verica .
c
1/2
1
=
c
1
x
2
+ c
1
x
2
2xc
1/2
1
x
=
2c
1
x
2
2c
1/2
1
x
2
. (3.88)
Por lo tanto podemos tomar las sigientes variables
t = x, u(t) =
y
x
, (3.89)
obteniendo as la siguiente ODE
u
t =
1
2
_
1
u
+ u
_
u, = u
2
1 =
C
1
t
, (3.90)
que en las variables originales
_
y
x
_
2
1 =
C
1
x
, (3.91)
es la solucin anterior.
3. Solucin mediante grupos de Lie.
La solucin:
y =
_
x
2
+ x _C1,
Clculo de las simetras. Tenemos que resolver la siguiente ecuacin en derivadas parciales:
x
+
1
2
(
y
x
) (x
2
+ y
2
)
x y

1
4
y
(x
2
+ y
2
)
2
x
2
y
2
(x, y) (
1
y

x
2
+ y
2
2 x
2
y
) (x, y) (
1
x

x
2
+ y
2
2 x y
2
) = 0
las posibles soluciones a esta ecuacin son:
_
= 0, =
x
2 y
_
,
_
= 0, =
x
2
2 y
+
y
2
_
, [ = x, = y] ,
_
= 1, =
y
2 x
_
,
_
= x
2
+ y
2
, (x, y) = 2xy
_
Por ejemplo si seguimos la tctica de los invariantes encontramos que el campo X = x
x
+ y
y
genera la
siguiente solucin::
dx
=
dy
=
dx
x
=
dy
y
= ln x = ln y = y = ax
esta sera la solucin que nos sugiere precisamente el uso directo del teorema Pi. Vemos que slo obten-
emos una solucin particular, pero que sta es invariante, de hecho si pensamos en la ode como un sistema
dinmico vemos que el punto jo de dicho sistema sera precisamente la solucin y = x.
De esta forma vemos el alcance de cada mtodo. Veamos ahora la solucin que sugiere las variables
cannicas (v.c.).
Las variables cannicas que obtenemos con el campo X =
x
+
x
2y
y
son:
_
t =
y
x
, u(t) = x
_
=
_
x = u(t), y = t
_
u(t)
_
por lo tanto la ode original queda expresada en estas nuevas variables como:
u
= 2t = u(t) = t
2
deshacemos el c.v. obteniendo
x =
y
2
x
+ C
1
que es precisamente la solucin que hemos ido obteniendo hasta ahora. Veamos por ltimo las variables
cannicas que induce la simetra X = x
x
+ y
y
,
_
t =
y
x
, u(t) = ln x
_
=
_
x = e
u(t)
, y = te
u(t)
_
por lo tanto la ode original queda expresada en estas nuevas variables como:
u
=
2t
t
2
1
= u(t) = ln (1 + t) ln (1 + t) + C
1
deshacemos el c.v. obteniendo
ln x = ln
_
y
x
1
_
ln
_
y
x
+ 1
_
+ C
1
ln x = C
1
+ ln
_
x
(y x) (y + x)
_
.
El cv que induce la simetra X =
x
2y
y
, es parecido al obtenido mediante el MD i.e. t = x, u(t) =
y
2
x
.
Ejemplo 3.5.2 Resolver la ecuacin:
_
y + xy
_
2
= 4x
2
y
(3.92)
Solucin. Atacaremos el problema por tres mtodos:
u
=
2(1 u
1 u)
x
=
du
(1 u
1 u)
=
2dx
x
efectuando el c.v.1 u = z, entonces:
_
dz
z
z
=
_
2dx
x
= ln |z 1| ln
z 1
z + 1
= ln
c
x
2
deshaciendo todos los c.v. obtenemos:
y = 2C
C
2
x
2. Solucin mediante AD.
es una ecuacin no lineal que no es homognea dimensionalmente si se discrimina. S lo ser si introduci-
mos una constante [a] = L
1
x
L
y
_
y + xy
_
2
= 4ax
2
y
(3.93)
Con lo que, desde el A.D., podemos plantear un cambio de variable a travs del monomio, eviden-
temente adimensional =
y
ax
del que obtenemos y = ax , y
= a + ax
que podemos sustituir en (3.93),

resultando
_
2 + x
_
2
= 4
_
+ x
_
(3.94)
Resolviendo la ecuacin de segundo grado en
que hemos obtenido
=
2
x
_
(1 )
_
1
_
(3.95)
y una vez ms se obtiene la separacin de variables, que permite la integracin:
d (1 )
(1 )
1
=
2
x
dx =
dz
z
z
= ln
c
x
2
(3.96)
ln |z 1| ln
z 1
z + 1
= ln
c
x
2
(3.97)
En denitiva se obtienen 2 soluciones con sentido fsico, para los dos signos,
c
x
2
=
_
z 1
_
2
z =
_
1 +
c
x
_
2
, = 1
_
1 +
C
x
_
2
= y = 2aC
aC
2
x
(3.98)
c
x
2
=
_
z + 1
_
2
z =
_
c
x
1
_
2
, = 1
_
C
x
1
_
2
= y = 2aC
aC
2
x
(3.99)
Las soluciones puramente matemticas son
y = 2C
C
2
x
. (3.100)
Vemos la alternativa pedestre:
_
y + xy
_
2
= 4ax
2
y
esta ecuacin es d.h. sii [a] = x

1
y, por lo tanto la solucin que sugiere el AD (mtodo pedestre(m.p.)) es:
y a x
x 0 1 1
y 1 1 0
= y = ax,
comprobamos que se trata de una solucin, como y
= 2ax entonces:
(ax + xa)
2
= 4ax
2
(a)
por lo tanto es solucin. Este monomio induce el siguiente cv.
t = x, u(t) =
y
x
,
entonces
_
2u + tu
_
2
= 4
_
u + tu
_
,
i.e. lo mismo pero ms sencillo y claro.
3. Mtodo de Lie.
La ecuacin es homogenea ,Dalembert .Tenemos que resolver la siguiente ecuacin
x
+
(
y
x
) (2 x y 2
_
x
2
y x)
x

y
(2 x y 2
_
x
2
y x)
2
x
2
(x, y)
_
_
_
2
2 xy
x
2
y x
x

2x y 2
_
x
2
yx
x
2
_
_
_
(x, y) (1 +
x
x
2
yx
)
x
= 0
[ = 0, =
_
x (x y)
x
], [ = 0, _ = x y +
_
x (x y)],
[ = 0, = 2 x
2
+ 2 y x 2
_
x (x y) x + y
_
x (x y)],
[ = x, = y], [ = 1, = 2
y
x
],
[ = x
2
(4 x 3 y), = y
2
x], [ = x (2 x y), = y
2
]
y la soluciones son:
y = x, y =
C
1
(C
1
+ 2 x)
x
, y =
C
1
(C
1
+ 2x)
x
Vemos que una de las posibles simetras de la ecuacin son las de escala generadas por [ = x, = y], esta
simetra genera la solucin que hemos obtenido mediante AD (m.p.). Vemos las diferencias entre ambos
mtodos.
Las variables cannicas nos llevan a los siguientes resultados: El campo X =
x (xy)
x

y
induce el siguiente
cambio de variables;
_
t = x, u(t) = 2
_
x (x y)
_
=
_
x = t, y =
1
4
4t
2
u(t)
2
t
_
la ode queda:
u
= 2 = u(t) = 2t + C
1
si deshacemos el c.v. obtenemos
2
_
x (x y) = 2x + C
1
Ejemplo 3.5.3 Estudiar la ecuacin
y
=
y
x
+
_
y
2
x
2
1 (3.101)
Solucin.
u
u
2
1
x
=
du
u
2
1
=
dx
x
cuya solucin es:
ln x ln(u +
_
u
2
1) + C
1
= 0
si expresamos esta solucin en funcin de las variables originales, entonces obtenemos:
x = C
1
_
y
2
(x)
x
2
1
_
+
C
1
x
y (x) .
Esta ode no ser d.h. si entendemos que el nmero 1 dentro de la ecuacin es precisamente un nmero
puro. Con la experiencia anterior podemos darle las dimensiones de una constante [A] = L
2
x
L
2
y
, obtener el
monomio adimensional =
x
2
A
y
2
, y de nuevo hacer el cambio de la variable dependiente y de su derivada:
y = x
_
A
_
1/2
, y
=
_
A
_
1/2
1
2
A
1/2
3/2

(3.102)
sustituimos y simplicamos
_
A
_
1/2
x
2
A
1/2
1/2
=
_
A
_
1/2
+
A = (3.103)
x
2
=
_
1 =
d
1
= 2
dx
x
(3.104)
que integrando
ln
1 1
1 + 1
= ln
1
x
2
+ ln
1
c
2
(3.105)
es decir

1 1
1 + 1
=
1
c
2
x
2
(3.106)
y obtenemos la solucin de (3.101):
= 1
_
1 x
2
c
2
1 + x
2
c
2
_
, (3.107)
y por lo tanto:
y =
_
A
2c
+
x
2
c
A
2
_
i.e. y =
_
1
2c
+
x
2
c
2
_
(3.108)
El m.p. en este caso nos conduce a la siguiente solucin particular:
y a x
x 0 2 1
y 1 2
= y
2
= ax
2
= y = a
x,
se comprueba con facilicidad que esta solucin particular verica la ecuacin diferencial original. El c.v.
que induce este monomio es el siguiente:
t = x, u(t) =
y
x
,
y por lo tanto obtenemos la siguiente ODE
u
t =
_
u
2
1, = ln x ln
_
y
x
+
_
y
2
x
2
1
_
+ c
1
= 0.
3. Solucin mediante GL.
La ecuacion que debemos resolver es la siguiente:
x
+ (
y
x
)(
y
x
+
_
y
2
x
2
1)
y
_
y
x
+
_
y
2
x
2
1
_
2
(x, y)
_
_
y
x
2

y
2
_
y
2
x
2
1x
3
_
_
(x, y)
_
_
1
x
+
y
_
y
2
x
2
1 x
2
_
_
= 0
donde las posibles solucione son:
[ = 0, =
_
y
2
x
2
1x], [ = 0, = x
2
+ y
2
_
y
2
x
2
1 yx],
[ = x, = y], simetras de escala
[ = x y, = x
2
2 y
2
], [ = x (x
2
2 y
2
), = y (3 x
2
4 y
2
)]
la solucion que induce la simetra X = x
x
+ y
y
a trasvs de las variables cannicas es:
_
t =
y
x
, u(t) = ln(x)
_
=
_
x = e
u(t)
, y = te
u(t)
_
u
=
1
t
2
1
= u(t) = ln
_
t +
_
t
2
1
_
+ C
1
ln(x) = ln
_
y
x
+
_
y
2
x
2
1
_
+ C
1
. (3.109)
De esta forma vemos las ventajas y desventajas de cada mtodo.
3.6 Ecuacin lineal de primer orden
Tiene la forma
dy
dx
+ P (x) y = Q(x) (3.110)
3.6. ECUACIN LINEAL DE PRIMER ORDEN 47
Separando variables
y = u(x)v(x) (3.111)
que nos lleva a
u
v + u
_
Pv + v
_
= Q (3.112)
La solucin por lo tanto es:
y =
_
_
Qe
_
Pdx
+ C
1
_
e
_
Pdx
(3.113)
La ODE ser discriminalmente homognea (d.h.) si
[P (x)] =
_
x
1
_
, [Q(x)] =
_
yx
1
_
(3.114)
si descomponemos y (x) en el producto y (x) = u (x) v (x) tendremos libertad para forzar a una de ellas, p. e.
u (x) , a que satisfaga alguna condicin, que convendr sea d.h.. As plantearemos una relacin entre u,
du
dx
,
P (x) que no puede ser otra (3.115) que
u
du
dx
P(x)
x 0 1 1
u 1 1 0
(3.115)
du
dx
[P (x) u]
1
= Cte (3.116)
lo ms sencillo ser hacer la constante igual a 1, resultando
du
dx
+ P (x) u = 0 (3.117)
Teniendo en cuenta que
dy
dx
= u
dv
dx
+ v
du
dx
(3.118)
sustituiremos en (3.114) obteniendo
v
_
du
dx
+ uP
_
+ u
dv
dt
= Q (3.119)
y teniendo en cuenta (3.116) resultar
u
dv
dx
= Q (3.120)
Tanto (3.116) como (3.117) permiten la integracin por separacin de variables obtenindose u (x), v (x) y en
consecuencia la solucin y (x) = uv de la (3.114).
3.6.3 Mtodo de Lie.
El sistema que se gebera es el siguiente:
x
+ (
y
x
) (Py + Q)
y
(Py + Q)
2
(x, y) (P
y + Q
) (x, y)(P) = 0
la simetra que se genera es:
_
= 0, = e
_
Pdx
_
esta simetra conduce al siguiente cambio de variables:
_
t = x, u(t) =
y
e
_
Pdx
_
=
_
x = t, y =
u(t)
e
_
P(t)dt
_
por lo tanto la ode original se escribe ahora en estas nuevas varibles como:
u
=
Q(t)
e
_
P(t)dt
= u(t) =
_
Q(t)e
_
P(t)dt
dt + C
1
reescribiendo esta solucin en las variables originales obtenemos que:
y =
_
_
Qe
_
Pdx
+ C
1
_
e
_
Pdx
3.6.4 Ejemplos.
Ejemplo 3.6.1 Sea de nuevo la ecuacin
y
= 3
y
x
+ x.
Solucin. Como en ejemplos anteriores estudiaremos este ejemplo empleandovarias tcticas
Reescribimos la ecuacin, que no es d.h., pero s lo ser si la ponemos en la forma
dy
dx
3
y
x
= Ax (3.121)
donde a la constante que introducimos, A, la suponemos dimensiones [A] = L
2
x
L
y
. Es inmediato que
con ella podemos construir el monomio adimensional = A
x
2
y
estableciendo la variable dependiente y su
derivada
y = A
x
2
dy
dx
=
2Ax Ax
2
2
.
Sustituidas en (3.121) resulta
2Ax Ax
2
2
3A
x
= Ax
Reordenando y simplicando trminos obtenemos
=
1
x
_
+
2
_
(3.122)
donde es posible la separacin de variables e integracin inmediata:
d
( + 1)
=
dx
x
= ln

+ 1
= ln
c
x
=
cAx
2
y

x
3
A
y
+ c = 0 = y = Ax
2
+
A
c
x
3
(3.123)
que sera la solucin de (3.121), correspondiente a un eventual problema fsico, representado por ella. Para
la solucin de (3.121) basta sustituir A por su valor numrico
y = x
2
+ Cx
3
(3.124)
donde C es una constante indeterminada. Esta solucin no corresponde a ningn problema fsico,
evidentemente.
Tambin podemos seguir este otro mtodo dimensional alternativo al anterior. Reescribimos la ecuacin
(3.75) como
y
=
3y
x
+ Ax (3.125)
donde A es cierta constante con dimensiones [A] = yx
2
. De esta forma intentamos encontar una solucin
aplicando directamente el teorema Pi a las magnitudes (y, x, c
1
) , obteniendo:
y x A
x 0 1 2
y 1 0 1
= y = Ax
2
(3.126)
viendo de esta forma que hemos obtenido una solucin particular a nuestra ecuacin diferencial por un
mtodo pedestre. El cv que obetemos es:
t = x, u(t) =
y
x
2
,
y por lo tanto
u
=
u + 1
t
, u = 1 + C
1
t,
y
x
2
= 1 + C
1
x.
llegando as al mismo resultado pero de forma mucho ms sencilla.
En primer lugar decimos que la ecuacin es de tipo lineal y que la solucin formal es:
y = x
2
+ C
1
x
3
Tenemos que resolver la ecuacion:
x
+ (
y
x
) (
3 y
x
+ x)
y
(
3 y
x
+ x)
2
(x, y) (
3 y
x
2
+ 1)
3 (x, y)
x
= 0
cuya solucion es:
[ = 0, = x
3
], [ = 0, = y + x
2
], [ = x, = 2 y],
[ = x
2
, = 3x y], [ = 3 y + x
2
, =
9 y
2
x
]
vemos que la simetra [ = x, = 2 y], induce por le mtodo de los invariantes i.e.
dx
x
=
dy
2y
= y = x
2
,
la misma solucin que la obtenida aplicando el teorema Pi.
Las variables cannicas obtenidas por el campo X = x
2
x
+ 3x y
y
son:
_
t =
y
x
3
, u(t) =
1
x
_
=
_
x =
1
u(t)
, y =
t
u(t)
3
_
,
con estas variables la ode se escribe como:
u
= 1 = u(t) = t + C
1
,
por lo tanto la solucin es:
1
x
=
y
x
3
+ C
1
= y = x
2
+ x
3
C
1
.
Ejemplo 3.6.2 Sea la ecuacin
_
1 x
2
_
dy
dx
+ xy = 1 (3.127)
Solucin. Atacaremos este ejemplo mediante dos tcticas
Se comprueba fcilmente que no es d.h.. Para conseguir la homogeneidad dimensional debemos introducir
dos constantes dimensionadas [A] = L
2
x
y [B] = L
x
L
y
, obteniendo
_
A x
2
_
dy
dx
+ xy = B (3.128)
De la resolucin de la matriz (3.129)
x y A B
L
x
1 0 2 1
L
y
0 1 0 1
1 1 1 1
2 0 1 0
(3.129)
obtenemos las dos variables adimensionales y , pudindose escribir, de acuerdo con el A. D.
=
A
B
y
x
= f
_
x
2
A
_
= f () = y =
B
A
x f () (3.130)
por lo tanto
y
=
B
A
f +
B
A
x
d f
d
d
dx
=
B
A
_
f + 2 f
_
(3.131)
Sustituyendo en (3.128)
_
A x
2
_
B
A
_
f + 2 f
_
+
B
A
x
2
f = B (3.132)
Reordenando, con las simplicaciones oportunas, obtenemos la posibilidad de separar variables y su inte-
gracin. En sntesis:
2 (1 ) f
= 1 f =
2d f
1 f
=
d
(1 )
(3.133)
f = 1 c
= y =
B
A
x c
B
A
_
A x
2
(3.134)
que ser la solucin de (3.128). Con A = 1, B = 1 , obtenemos la solucin de (3.127)
y = x c
_
1 x
2
(3.135)
donde c ser una constante arbitraria.
Vemos que hemos obtenido dos monomios pero que ninguno de ellos es solucin particular, por lo que
deberemos buscar una combinacin que nos de una solucin particular, sta es:
1
=
x
2
A
,
2
=
yx
B
,
2
1
1
=
y
x
,
de la que obtenemos el siguiente c.v.
t = x, u(t) =
y
x
,
u + tu
=
tu
2
1
t
2
1
, = u = 1 +
_
(t + 1) (t 1)
t
y por lo tanto
y
x
= 1 +
_
(x + 1) (x 1)
x
.
La ecuacin es lineal y la solucion formal es:
y = x c
x 1
x + 1 (3.136)
Tenemos que resolver la siguiente ecuacion en derivadas parciales:
x
+
(
y
x
) (1 x y)
1 x
2

y
(1 x y)
2
(1 x
2
)
2
(x, y) (
y
1 x
2
+
2 (1 x y) x
(1 x
2
)
2
) +
(x, y) x
1 x
2
= 0, (3.137)
_
= 0, =
_
1 + x
2
_
,
_
= 0, =
_
(x 1) (x + 1)
_
,
[ = 0, = y + x] ,
_
= 0, =
(2xy y
2
1)(y + x)
(x 1)/(x + 1)
_
(3.138)
que generan las siguientes soluciones:
Por ejemplo con
_
= 0, =
1 + x
2
_
obtenemos:
_
t = x, u(t) =
y
1 + x
2
_
=
_
x = t, y = u(t)
_
1 + t
2
_
la ode se escribe ahora como:
u
1 + t
2
1 2t
2
+ t
4
cuya solucin es:
u(t) =
_
1 + t
2
_
3/2
4 (t + 1)
2
+
_
1 + t
2
_
3/2
4 (t 1)
2
+ C
1
por lo tanto la solucin en las variables originales resultar:
y
x
2
1
=
C
1
(x
2
1) + x
x
2
1
x
2
1
= y = x c
x 1
x + 1.
Otro cambio de variables lo podemos obtener de [ = 0, = y + x] de tal forma que:
(t = x, u(t) = ln(y + x)) =
_
x = t, y =
te
u(t)
1
e
u(t)
_
la ode se escribe ahora como:
u
=
t
1 + t
2
cuya solucin es:
u(t) =
1
2
ln(t 1)
1
2
ln(t + 1) + C
1
por lo tanto la solucin en las variables originales resultar:
ln(y + x) =
1
2
ln(x 1)
1
2
ln(x + 1) + C
1
slo queda simplicar esta expresin, obtenindose:
ln
1
y + x
= C
1
+ ln
1
_
(x 1)
_
(x + 1)
y = x c
x 1
x + 1
como ya sabamos.
3.7 Ecuaciones exactas.
Llamaremos exacta a una ODE del tipo
P(x, y(x))dx + Q(x, y(x))dy = 0, i.e. y
=
P(x, y(x))
Q(x, y(x))
(3.139)
donde
D
2
(P)(x, y(x)) = D
1
(Q)(x, y(x)). (3.140)
La forma general de resolverlas es la siguiente. Queremos encontrar una funcin F(x, y) tal dF = Pdx + Qdy
(1-forma de Pfafft)
_
x
P(x, y)dx +
_
y
Q(x
0
, y)dy = C.
Entonces la solucin ser F(x, y) = C
1
i.e.
F(x, y(x)) + C
1
= 0, (3.141)
dicha funcin siempre existe si se verica la condicin (3.140). Podemos reescribir la ecuacin (3.139) como
sigue:
x
F +
y
F
_
y
_
= 0 (3.142)
donde obviamente
x
F = P y
y
F = Q.
Lo mejor es que veamos algn ejemplo.
Ejemplo 3.7.1 Integrar la ecuacin exacta
(sin(xy) + xy cos(xy)) dx + x
2
cos(xy)dy = 0 (3.143)
Solucin. Veremos como siempre distintas tcticas
1. Mtodo tradicional. Lo primero es comprobar que realmente es una ode exacta i.e.
y
P =
x
Q
viendo de forma inmediata que:
y
P =
y
(sin(xy) + xy cos(xy)) = 2 (cos xy) x x
2
y sin xy
x
Q =
x
_
x
2
cos(xy)
_
= 2 (cos xy) x x
2
y sin xy
3.7. ECUACIONES EXACTAS. 53
por lo tanto la ode dada es exacta. Calculamos la funcin F(x, y) como sigue:
F(x, y) =
_
x
(sin(xy) + xy cos(xy)) dx +
_
y
x
2
0
cos(x
0
y)dy
(sin xy) x = C
1
2. Grupos de Lie. En primer lugar calculamos la ecuacin con la que podremso obtener los innitesimales
(
y
x
) (sin(x y) + x y cos(x y))
x
2
cos(x y)

y
(sin(x y) + x y cos(x y))
2
x
4
cos(x y)
2

(
2 cos(x y) y x y
2
sin(x y)
x
2
cos(x y)
+
2 (sin(x y) + x y cos(x y))
x
3
cos(x y)

(sin(x y) + x y cos(x y)) (x y) y
x
2
cos(x y)
2
)
(
2 cos(x y) x x
2
y sin(x y)
x
2
cos(x y)

(sin(x y) + x y cos(x y)) sin(x y)
x cos(x y)
2
) = 0
de esta forma obtenemos que las posibles soluciones a esta ecuacin son:
[ = 0, =
1
x
2
cos(x y)
], [ = 0, =
1 + cos(2x y)
cos(x y)
]
[ = 0, =
sin(x y)
x cos(x y)
], [ = x, = y]
de esta forma vemos que por ejemplo el campo X = x
x
+ y
y
induce el siguiente cambio de varibles:
(t = xy, u(t) = ln x) =
_
x =
1
e
u(t)
, y = te
u(t)
_
por lo tanto la ecuacin (3.143) se reescribe como:
u
=
cos t
sin t
= u(t) = ln(sin t) + C
1
deshaciendo el cambio de variables obtenemos:
ln x = ln(sin(xy)) + C
1
= sin(xy)x = C
1
El mtodo de los invariantes nos llevara a buscar una solucin del tipo
dx
x
=
dy
y
= yx = C
1
que obviamente no es solucin, cualquiera de las otras simetra conducira a una solucin del tipo y =
const.
Algunas ecuaciones no verican la relacin
y
P =
x
Q, sin embargo se puede intentar buscar cierta funcin
(factor integrante) (x, y) que haga que la ecuacin
(P(x, y)dx + Q(x, y)dy) = 0
sea exacta. Dicho factor lo deberemos buscar de la siguiente manera.
Como
y
(P) =
x
(Q) ,
entonces
Q
x
() P
y
() =
_
y
P
x
Q
_
,
por lo tanto
Q
x
ln P
y
ln =
y
P
x
Q,
observndose que si
y
P
x
Q
Q
= h(x) = (x) = e
(
_
h(x)dx)
,
mientras que si
y
P +
x
Q
P
= h(y) = (y) = e
(
_
h(y)dy)
.
El siguiente teorema estable la relacin entre el factor integrante y las simetras de una ode de primer orden.
Teorema 3.7.1 (Factor integrante de Lie) . La ecuacin P(x, y)dx +Q(x, y)dy = 0 admite un grupo uniparametrco
G cuyo generador innitesimal X es tal X =
x
+
y
sii
=
1
P + Q
, P + Q = 0, (3.144)
es un factor integrante de la ecuacin dada.
Veamos el siguiente ejemplo.
Ejemplo 3.7.2 Integrar la ecuacin:
_
x + y
2
_
dx 2xydy = 0
Solucin. Como siempre veremos al menos dos tcticas:
1. Mtodo tradicional. Vemos que
y
P =
x
Q ya que
y
P = 2y y
x
Q = 2y
como en nuestro caso se tiene que
y
P
x
Q
Q
=
1
2x
= h(x) = (x) = e
(
_
1
2x
dx)
=
1
x
2
entonces
1
x
2
__
x + y
2
_
dx 2xydy
_
= 0
es exacta y su solucin es:
x = C
1
e
y
2
2
.
2. Mtodo de Lie. En primer lugar calcularemos la ecuacin
x
+
1
2
(
y
x
)(x + y
2
)
x y

1
4
y
(x + y
2
)
2
x
2
y
2
(
1
2 x y

x + y
2
2 x
2
y
) (
1
x

x + y
2
2 x y
2
) = 0
3.8. ECUACIN DE BERNOULLI 55
[ = 0, =
x
2 y
], [ =
x
2
x + y
2
, = 0], [ = 2x, = y]
si por ejemplo nos jamos en la simetra X = 2x
x
+ y
y
, vemos que nos conduce al siguiente cambio de
variables (variables cannicas);
_
t =
y
x
, u(t) =
1
2
ln x
_
=
_
x = e
2u(t)
, y = t
_
e
2u(t)
_
la ode original se escribe en estas variables como
u
= t = u(t) =
t
2
2
+ C
1
que en las variables originales se escribe como:
1
2
ln x =
y
2
2x
+ C
1
= x = C
1
e
y
2
2
.
Si atendemos al teorema del factor integrante de Lie, vemos que con [ =
x
2
x+y
2
, = 0] obtenemos
=
1
P + Q
=
1
x
2
x+y
2
(x + y
2
)
=
1
x
2
tal y como habamos calculado antes.
Observamos que la solucin que induce la simetra X = 2x
x
+ y
y
por el mtodo de los invariantes es:
y = C
1
x
que NO es solucin de la ode dada.
3.8 Ecuacin de Bernoulli
Se escribe
y
+ P (x) y = Q(x) y
n
(3.145)
o alternativamente:
y
= f
n
(x)y
n
+ f
1
(x)y (3.146)
Se pueden seguir varias tcticas como las siguientes:
El cambio de variable
y(x) = u(t)
(
1
1n
)
, x = t (3.147)
que conduce a la siguiente ecuacin diferencial
u
(t) =
_
u(t)P(t) + Q(t)
_
u(t)
(
1
1n
)
_
n
u(t)
(
n
1n
)
_
(n 1) (3.148)
que es una ode de tipo lineal.
Un segundo mtodo sugiere hacer
y(x) = u(x)v(x)
variables separadas. Se iguala a cero el coeciente en u (v
+ f
1
v = 0) determinando as v, apareciendo ahora
una ecuacin en u(x) de variables separadas.
La solucin queneral de una ode tipo Bernoulli viene dada por el siguiente formuln:
y =
e
(
1
n1
_
P(x)dx)
_
n
_
e
_
P(x)dx
Q(x)
e
(n
_
P(x)dx)
dx +
_
e
_
P(x)dx
Q(x)
e
n
_
P(x)dx
dx + C
1
_ 1
n1
e
n
_
P(x)dxn
n1
(3.149)
o de forma algo ms compacta:
y
1n
=
(1 n)
_
Q(x)e
(n1
_
P(x)dx)
+ C
1
e
(n1
_
P(x)dx)
(3.150)
Para su estudio dimensional conviene dividir por el segundo miembro, resultando
1
Qy
n
dy
dx
+
P
Qy
n1
= 1 (3.151)
Las dimensiones de
P
Q
sern
_
P
Q
_
=
_
y
n1
_
(3.152)
y para que el segundo trmino resulte adimensional proceder tomar como nueva variable
t = y
1n
(3.153)
si tambin se desea que sea lineal en t.
Para eliminar
dy
dx
derivaremos (3.153) respecto x
dt
dx
=
1 n
y
n
dy
dx
(3.154)
y de esta ecuacin y la (3.153) resultar
1
(n 1)
dt
dx
+ P (x) t = Q(x)
que es una ecuacin de primer orden lineal en t, de cuya solucin hemos tratado en el apartado anterior.
3.8.3 Mtodo de Lie.
Tenemos que resolver la siguiente ecuacin etc...
x
+ (
y
x
) (p(x) y + q(x) y
n
)
y
(p(x) y + q(x) y
n
)
2
(x, y) (p
y + q
y
n
)
(x, y) (p(x) +
q(x) y
n
n
y
) = 0
cuyas posibles soluciones son:
[ = 0, = e
(
_
p(x) (1+n) dx)
y
n
],
_
=
e
(
_
p(x) dx (1+n))
q(x)
, =
yp(x) e
(
_
p(x) dx (1+n))
q(x)
_
el cambio de variables que induce la primera de las simetras es:
_
u(t) =
y
(1n)
(1 n)e
((1+n)
_
p(x) dx)
, t = x
_
por lo tanto la transformacin inversa es:
_
_
y = e
_
ln(u(t) e
(
_
p(t) dt+n
_
p(t) dt)
u(t) e
(
_
p(t) dt+n
_
p(t) dt)
n)
1+n
_
, x = t
_
_
la ecuacin original queda por lo tanto:
u
t
=
q(t)
_
(u(t) e
(
_
p(t) dt (1+n))
(1 + n))
(
1
1+n
)
_
n
(1 + n) u(t)
(u(t) e
(
_
p(t) dt (1+n))
(1 + n))
(
1
1+n
)
cuya solucin es:
u(t) C
1
_
t
q(b)
_
(u(t)e
((n1)
_
p(b) db )
(n 1))
(
1
n1
)
_
n
(n 1) u(t)
_
u(t)e
((n1)
_
p(b)db)
(n 1)
_
(
1
1+n
)
db = 0
deshaciendo el cambio de variables obtenemos por lo tanto nuestro formuln (3.150):
Tambin podemos probar a calcular el factor integrante de Lie siendo ste:
:= e
(
_
p(x) dx (1+n))
y
(n)
y por lo tanto la solucin de la ode exacta resultante es:
e
(
_
p(x) dx (1+n))
y
(1n)
1 n

_
q(x) e
(
_
p(x) dxn
_
p(x) dx)
dx + C
1
= 0
Veamos algunos ejemplos.
3.8.4 Ejemplos.
xy
+ y = y
2
ln x
Solucin. En este ejemplo veremos dos tctics tanto la tradicional como la de Lie, ya que el mtodo dimensional
no funciona por haber funciones del tipo (ln x) .
1. Mtodo tradicional nos lleva a efectuar el siguiente cambio de variables y = uv de donde
xu
v + u
_
xv
+ v
_
= u
2
v
2
ln x
resolviendo
_
xv
+ v
_
= 0 = v =
1
x
entonces
u
=
u
2
x
2
ln x =
1
u
=
ln x
x

1
x
C
1
por lo tanto
u =
x
1 + C
1
x + ln x
y =
1
1 + C
1
x + ln x
2. Con el mtodo de Lie tendremos que resolver la siguiente ecuacin en derivadas parciales (edp):
x
+
(
y
x
) (y + y
2
ln(x))
x

y
(y + y
2
ln(x))
2
x
2
(
y
2
x
2

y + y
2
ln(x)
x
2
)
(1 + 2 y ln(x))
x
= 0
[ = 0, = y
2
x], [ = 0, = y (1 + y ln(x) + y)],
[ = 0, =
(1 + y ln(x) + y)
2
x
], [ = x, = y
2
]
la primera de las simetrias nos lleva a obtener las siguientes v.c.
_
t = x, u(t) =
1
y x
_
=
_
x = t, y =
1
u(t) t
_
la ode original queda en estas variables como:
u
t
=
ln(t)
t
2
= u(t) =
ln(t)
t

1
t
+ C
1
que en las variables originales queda
1
y x
=
ln(x)
x

1
x
+ C
1
tal y como queramos hacer ver.
y
=
y
x + 1

(x + 1)
3
y
2
2
Solucin. En este caso slo utilizaremos la tctica de lso grupos de Lie.
1. El mtodo de Lie nos lleva a la siguiente edp
x
+ (
y
x
) (
y
x + 1

(x + 1)
3
y
2
2
)
y
(
y
x + 1

(x + 1)
3
y
2
2
)
2
(
y
(x + 1)
2

3 (x + 1)
2
y
2
2
) (
1
x + 1
(x + 1)
3
y) = 0
[ = 0, =
(x + 1) y
2
2
], [ = 0, =
(4 y x
3
+ 6 y x
2
+ 3 y x + y x
4
6)
2
x + 1
]
[ = 0, =
(y x
4
+ 4 y x
3
+ 6 y x
2
6 y x 6 9 y) y
2
], [ = x + 1, = 4 y]
por lo tanto el mtodo de las variables cannicas nos conduce a (X =
(x+1) y
2
2

y
):
_
t = x, u(t) =
2
y (x + 1)
_
=
_
y =
2
u(t) (t + 1)
, x = t
_
u
t
= t
2
2 t 1 = u(t) =
1
3
t
3
t
2
t + C
1
la solucin en las variables originales resulta:
2
y (x + 1)
=
1
3
x
3
x
2
x + C
1
,
tal y como queramos ver.
2. Mtodo dimensional. Reescribiendo la ODE de la siguiente forma:
y
=
y
x + a

b (x + a)
3
y
2
2
,
donde claramente, [a] = x, [b] = x
4
y
1
, podemos probar con el siguiente cv (observando adems la forma
de la ODE)
t = x, u(t) = (x + 1)
4
y,
donde (x + 1)
4
y = 1, es solucin particular de la ODE. De esta forma obtenemos
(t + 1) u
= 3u
1
2
u
2
, =
1
u
=
1
6
+
C
1
(t + 1)
3
,
en las variables originales tendremos
1
(x + 1)
4
y
=
1
6
+
C
1
(x + 1)
3
,
tal y como quermos hacer ver.
xy
+ xy
2
y = 0
Solucin. En este caso slo utilizaremos la tctica de los grupos de Lie.
1. El mtodo de Lie nos lleva a la siguiente edp
x
+ (
y
x
)
_
xy
2
+ y
x
_
y
_
y xy
2
x
_
2
_
y
2
x

y xy
2
x
2
_
_
2xy + 1
x
_
= 0
[ = y, = y
3
], [ = 0, =
y
2
x
]
[ = 0, = 2y + xy
2
], [ = x, = y]
[ = 0, = x (xy 2)
2
], [ =
1
y
x, =
1
x
]
por lo tanto el mtodo de las variables cannicas nos conduce a (X = x
x
y
y
):
(t = xy, u(t) = ln x) =
_
y =
t
e
u(t)
, x = e
u(t)
_
u
t
=
1
t (t 2)
= u(t) =
1
2
ln (t 2) +
1
2
ln (t) + C
1
la solucin en las variables originales resulta:
ln x =
1
2
ln (xy 2) +
1
2
ln (xy) + C
1
i.e.
y =
2x
x
2
+ 2C
1
,
tal y como queramos ver.
El factor integrante de la ode es:
=
1
y
2
Claim 3.8.1 La solucin por invariantes que induce el campo X = x
x
y
y
, es solucin de la ode dada, i.e.
y = 2x
1
.
2. El mtodo dimensional nos lleva a reescribir la ODE como
y
= y
2
+
y
x
,
por lo que slo habr que considerar una nica constante dimensional. El cv que suguiere el AD es el
siguiente
t = x, u(t) = xy,
donde xy = 1, es solucin particular de la ODE. De esta forma obtenemos
2u + tu
+ u
2
= 0, =
1
u
=
1
2
+
C
1
t
2
,
en las variables originales tendremos
1
xy
=
1
2
+
C
1
x
2
,
tal y como quermos hacer ver.
Ejemplo 3.8.4 Solve the following Bernoulli ode.
y
=
y
x + 1

(x + 1) y
2
2
. (3.155)
Observacin 3.8.1 Historically it was Bernoulli the rst person who introduced c.v. in order to solve odes (now bearing
his name). He managed to reduce this equation to a simpler linear equation.
Solution. It is observed that with
[a] = X, [b] = X
2
Y
1
, (3.156)
eq. (3.155) yields d.h. i.e.
y
=
y
x + a

b (x + a) y
2
2
, (3.157)
in this way it is obtained the following variables
_
t =
x
a
, u(t) =
1
abxy
_
=
_
x = at, y =
1
a
2
btu
_
, (3.158)
with this change of variables eq. (3.155) yields
2u + 2t (1 + t) u
= (t + 1)
2
, (3.159)
which is linear and its solution is:
u =
_
t
2
+ C
1
__
1 +
1
t
_
, (3.160)
and therefore in the original variables the solution yields
y =
2
(bx + C
1
) (x + a)
, (3.161)
in this case D.A. does not bring us to obtain a good change of variables but helps us to obtain a simpler ode than
the original one. We can try now with the following c.v.
_
t = x, u(t) =
1
x
2
y
_
=
_
x = t, y =
1
t
2
u
_
, (3.162)
which brings us to obtain a linear ode
t(t + 1)
_
2u
t + 4u
_
= 2t
2
u + t
2
+ 2t + 1, (3.163)
but with this tactic we are not able to obtain a simpler ode.
As we can see D.A. induces us to the c.v.
1
abxy
but y =
1
x
or y =
1
x
2
, are not particular solutions of (3.155).
Nevertheless solutions as y =
1
x + 1
or y =
1
(x + 1)
2
are particular and invariant solutions for this reason these
c.v. induce us to obtain an ode in separate variables, but unfortunately D.A. is unable to construct such c.v.. For
this reason we can only reduced our ode to a linear ode as in the theoretical case i.e. employing the theoretical
method purposed by Bernoulli.
Observacin 3.8.2 As in the above example if we consider the following cv then we get
t = x, u(t) = (x + 1)y,
where (x + 1)y = 1, is a particular solution. Thus we get
u
=
1
2
u
2
, =
1
u
=
t
2
+ C
1
,
in the original variables we obtain
1
(x + 1)y
=
x
2
+ C
1
,
as it is expected.
If we study this ode under the Lie group method, it is observed that eq. (3.155) admits the following symmetries
obtained from:
x
+ (
y
x
)
_
y
x + 1

(x + 1) y
2
2
_
y
_
y
x + 1

(x + 1) y
2
2
_
2
_
y
(x + 1)
2

1
2
y
2
_

_
y xy
1
x + 1
_
= 0 (3.164)
X
1
=
y
2
(x + 1)
2

y
, X
2
= (x + 1)
x
2y
y
, X
3
=
x
y
(x + 1)
y
, X
4
= x
x
(1 + 2x) y
(x + 1)

y
,
(3.165)
and which respective invariants are:
I
1
= x, I
2
= y (x + 1)
2
, I
3
= y (x + 1) , I
4
= yx (x + 1) ,
that induces the following canonical variables. For example from X
1
:
_
t = x, u(t) =
2
y (x + a)
_
=
_
x = t, y =
2
u (t + a)
_
, (3.166)
in such a way that eq. (3.155) yields
u
= b = y = bt + C
1
, (3.167)
and in the original variables i.e. (x, y) this solution yields
y =
2
(bx + C
1
) (x + a)
, (3.168)
as we already know.
3.9 Ecuaciones tipo Riccati.
Este tipo de ecuaciones tine la siguiente pinta:
y
= f
2
y
2
+ f
1
y + f
0
, (3.169)
donde las ( f
i
)
2
i=0
= f
i
(x) i.e. son funciones que slo dependen de la variable independiente x. Para resolverlas
por los mtodos tradicionales es necesario conocer una solucin particular y
p
(lo cual no siempre es fcil de
encontrar). Si ste es el caso entonces haciendo el cambio de funcin y = y
p
+ z entonces la ode (3.169) se
combierte en una de tipo Bernoulli con n = 2.
Se conocen las siguientes condiciones de integrabilidad:
1.
i
f
i
= 0
2. Existen (d
i
)
2
i=0
R
d
1
f
2
+ d
1
d
2
f
1
+ d
2
2
f
0
= 0
3. Sean (x), (x) funciones tales que
2
f
2
+ f
1
+ f
0
=
_
log

En general es bastante complicado obtener una solucin terica de este tipo de ecuaciones por eso slo veremos
unos cuantos ejemplos empleando exclusivamente el mtodo de Lie.
3.9. ECUACIONES TIPO RICCATI. 63
3.9.1 Ejemplos.
Ejemplo 3.9.1 Solve the following Riccati ode
y
= xy
2
2y
x

1
x
3
. (3.170)
Solution. It is observed that introducing the following dimensional constants eq. (3.170) veries the principle of
dimensional homogeneity (p.d.h.),
y
=
xy
2
a

2y
x

a
x
3
, (3.171)
where [a] = X
2
Y. As we can see if we need only one constant then our equation is scale invariant and therefore
this constant indicates us that the generator of this symmetry is X = 2x
x
+y
y
as we will see below. Therefore
we have the following variables
_
t = x, u(t) =
a
yx
2
_
=
_
x = t, y =
a
tu
2
_
, (3.172)
with this new variables eq. (3.170) yields
u
u
2
1
=
1
t
= ln t + arctanh(u) + C
1
= 0, (3.173)
and in the original variables this solutions is written as:
ln x + arctanh(
a
yx
2
) + C
1
= 0. (3.174)
We would like to emphasize that in this case, as we have only one dimensional constant, we can nd the partic-
ular solution y = ax
2
applying the Pi theorem, and that such solution (invariant particular solution) is induced
by the scale symmetry.
Eq. (3.170) has the following symmetries:
x
+ (
y
x
)(x y
2
2 y
x

1
x
3
)
y
(xy
2
2 y
x

1
x
3
)
2
(y
2
+
2 y
x
2
+
3
x
4
)
_
2xy
2
x
_
= 0 (3.175)
and hence
X
1
=
1
x
2
x
4
y
, X
2
= x
4
(y +
1
x
2
)
2
y
, , X
3
= x
x
2y
y
X
4
=
_
1
x
2
+ y
2
x
2
_
y
, X
5
= x
3
_
2 + 4y x
2
_
y
. (3.176)
and which corresponding invariants are:
I
1
=
yx
2
1
x
2
, I
2
= I
4
= x, I
3
= yx
2
, I
5
= yx
2
_
x
2
+ 1
_
. (3.177)
The method of canonical variables brings us to obtain the following ode, for example, for the symmetry X
5
, we
have
_
t = x
2
(y x
2
+ 1), u(t) =
1
2 x
2
_
, (3.178)
u
t
=
1
t
2
= u(t) =
1
t
+ C
1
, (3.179)
and hence
1
2 x
2
=
1
x
2
(y x
2
+ 1)
+ C
1
, (3.180)
as we already know.
The transformation induced by X
3
(scaling symmetry) is the following one:
_
t = y x
2
, u(t) = ln(x)
_
=
_
x = e
u(t)
, y =
t
(e
u(t)
)
2
_
, (3.181)
u
t
=
1
t
2
1
= u(t) = arctanh(t) + C
1
, (3.182)
therefore
ln(x) = arctanh(y x
2
) + C
1
, (3.183)
which looks a little different than the other solution.
Ejemplo 3.9.2 Solve the following Riccati ode:
y
=
y + 1
x
+
y
2
x
3
. (3.184)
Solution. The ode veries the p.d.h. if we introduce the following dimensional constant:
y
=
y + a
x
+ b
y
2
x
3
, (3.185)
where [a] = Y, [b] = X
2
Y
1
. In this way it is obtained the c.v.
_
t =
ax
2
by
2
, u(t) =
y
a
_
=
_
x =
abtu
2
, y = au
_
, (3.186)
and therefore eq. (3.184) yields
u
u
+ t =
ut
(ut + t + 1)
, (3.187)
which is a Bernoulli ode and its solution is:
1
u

t
_
arctan
t + C
1
_
= 0, (3.188)
hence in the original variables the solution is
a
y

ax
2
by
2
_
arctan
ax
2
by
2
+ C
1
_
= 0. (3.189)
It is observed that we have the particular solution y = Cx, obtained from the relationship
ax
2
by
2
.
Since this c.v. is not very good for our purposes we may now proceed as follows:
axy
= ay + 1 + b
2
y
2
x
2
(3.190)
where
[a] = y
1
, [b] = y
1
x (3.191)
3.9. ECUACIONES TIPO RICCATI. 65
in such a way that these constants induce the following change of variables:
_
t =
a
b
x, u(t) =
x
by
_
=
_
x =
b
a
x, y =
t
au(t)
_
(3.192)
which brings us to obtain
u
2
+ 1
=
1
t
2
=
1
t
+ arctan(u) + C
1
= 0 (3.193)
and hence
b
ax
+ arctan(
x
by
) + C
1
= 0 (3.194)
which is a better approximation. In the same way we can obtain a quadrature simply considering t =
b
ax
and
u(t) =
x
by
. in this way it is obtained u
= u
2
+ 1.
The Lie method bring us to solve the following pde
x
+ (
y
x
)(
y + 1
x
+
y
2
x
3
)
y
(
y + 1
x
+
y
2
x
3
)
2
(
y + 1
x
2

3 y
2
x
4
) (
1
x
+
2 y
x
3
) = 0, (3.195)
and which solutions are:
X
1
=
_
x +
y
2
x
_
y
, X
2
=
_
(ycos(
1
x
) + xsin(
1
x
))
2
x
_
y
X
3
=
_
sin(
1
x
)
2
(y sin(
1
x
) cos(
1
x
) x)
2
x (1 + cos(
1
x
)
2
)
_
y
, X
4
= x
2
x
+ xy
y
(3.196)
and which invariants are:
I
1
= I
2
= I
3
= x, I
4
=
y
x
. (3.197)
The symmetry X
1
=
_
x +
y
2
x
_
y
generates the following c.v.:
_
t = x, u(t) = arctan
_
y
x
__
= (y = tan(u(t)) t, x = t) , (3.198)
u
t
=
1
t
2
= u(t) =
1
t
+ C
1
, (3.199)
and hence the solution is
arctan
_
y
x
_
=
1
x
+ C
1
, (3.200)
as we already know.
We would like to emphasize that this ode is not scale invariant and nevertheless we have been able to obtain a
c.v. that reduces the ode to a quadrature.
Ejemplo 3.9.3 Resolver la ecuacin
y
=
(y x ln x)
2
x
2
+ ln x
Solucin. El mtodo de Lie nos lleva a resolver la siguiente edp
x
+ (
y
x
)(
(y x ln(x))
2
x
2
+ ln(x))
y
(
(y x ln(x))
2
x
2
+ ln(x))
2
_
2 (y x ln(x)) (ln(x) 1)
x
2

2 (y x ln(x))
2
x
3
+
1
x
_
2
(y x ln(x))
x
2
= 0
cuyas soluciones son
_
= 0, =
(y
1
2
(2 ln(x) + 1 +
5) x)
2
x
(
5)
x
_
,
_
= 0, =
x
(
5)
(y
1
2
(2 ln(x) + 1
5) x)
2
x
_
[ = 0, = x y +
y
2
x
2 y ln(x) + x ln(x)
2
+ x ln(x)], [ = x, = x + y]
el mtodo de las variables cannicas generadas por la simetra X = x
x
+ (x + y)
y
nos facilita el siguiente c.v.
_
t =
y x ln(x)
x
, u(t) = ln(x)
_
=
_
x = e
u(t)
, y = t e
u(t)
+ e
u(t)
u(t)
_
u
t
=
1
t 1 + t
2
= u(t) =
2
5
5 arctanh(
(2 t 1)
5
5
) + C
1
ln(x) =
2
5
5 arctanh(
1
5
(
2 (y x ln(x))
x
1)
5) + C
1
tal y como queramos hacer ver
3.10 Ecuaciones tipo Abel y Chini.
La expresin ms general de una ode tipo Abel tiene la siguiente pinta:
y
= f
3
y
3
+ f
2
y
2
+ f
1
y + f
0
(3.201)
mientras que la expresin ms general para una ode tipo Chini es:
y
= f
n
y
n
f
1
y + f
0
(3.202)
donde las funciones ( f
i
)
3
i=0
= f (x) son funcin slo de la variable independiente.
3.10.1 Mtodo tradicional
Hasta el momento (que yo sepa) no existe un mtodo para resolverlas aunque existen casos particulares que si
son susceptibles de ser atacados. Para estos casos el mtodo ms general hasta ahora es el denominado mtodo
de los invariantes de Abel (mtodo ste debido a Chini).
El invariante de una ode tipo Abel con f
2
= 0 es:
1
27
(f
0
f
3
+ f
0
f
3
+ 3f
0
f
3
f
1
)
3
f
4
3
f
5
0
(3.203)
Si este invariante no depende de x entonces la ode se puede resolver directamente. Para una ode Abel con
f
2
= 0, nos podemos cargar este trmino utilizando la siguiente transformacin
y = u(x)
1
3
f
2
f
3
. (3.204)
3.10. ECUACIONES TIPO ABEL Y CHINI. 67
Si f
0
= f
1
= 0 y
_
f
3
f
2
_
= a f
2
entonces en este caso se conoce la solucin. Como la ode se queda reducida a
y
= f
3
y
3
+ f
2
y
2
, (3.205)
entonces mediante el siguiente cambio de variables
x = t, y =
f
2
(t)r(t)
f
3
(t)
, (3.206)
y teniendo en cuenta la condicin
_
f
3
f
2
_
= a f
2
i.e.
f
3
f
2
f
2
f
3
f
2
2
= a f
2
, (3.207)
entonces la ecuacin (3.205) se transforma en una ode tipo variables separables
f
2
r
f
3
ra f
3
2
f
2
3
=
f
3
2
r
3
+ r
2
f
3
2
f
2
3
. (3.208)
tenemos por lo tnato la siguiente proposicin:.
Proposicin 3.10.1 Una ode tipo Abel de la forma
y
= f
3
y
3
+ f
2
y
2
puede ser integrada de forma exacta sii las funciones f
3
y f
2
satisfacen la relacin
_
f
3
f
2
_
= Kf
2
/ K = const.
Otro caso es el siguiente pero tiene la desventaja de depender del conocimiento previo de una solucin particular
y
1
adems de satisfacerse las siguiente relaciones:
Teorema 3.10.1 Si las funciones ( f
i
)
3
i=0
satisfacen la relacin
_
f
3
G(x)
_
= KG(x) / K = const.
tal que
G(x) =
_
f
2
2
3f
3
f
1
= 3f
3
y
1
+ f
2
donde y
1
es una solucin particular de la ode (3.201) entonces la solucin viene dada por la siguiente expresin:
y =
G
f
3
V +
_
f
2
G
3f
3
_
con
V =
_
f
3
3f
3
y
1
+ f
2
_
(y y
1
)
Encontramos las siguientes ecuaciones de tipo ABEL.
1. Segunda especie clase A.
En este caso la ode tiene la siguiente pinta:
(y + g(x)) y
= f
2
y
2
+ f
1
y + f
0
(3.209)
para este caso si
f
0
= f
1
g f
2
g
2
(3.210)
entonces existe solucin
y(x) = g(x) (3.211)
y(x) =
_
_
(f
2
g + f
1
) e
_
f
2
dx
dx + C
1
_
e
_
f
2
dx
(3.212)
tambin encontramos que si
f
1
= 2f
2
g g
(3.213)
la ode entonces se combierte en:
(y + g(x)) y
= f
2
y
2
+
_
2f
2
g g
_
y + f
0
(3.214)
para este caso se puede buscar o un factor integrante o en su defecto simetras
X =
e
2
_
f
2
dx
y + g

y
(3.215)
encontrndose una solucin explcita:
y =
1
2
_
_
_
_
2e
(2
_
f
2
dx)
g 2
_
e
(2
_
f
2
dx)
g
2
+ 2e
(2
_
f
2
dx)
_
_
_
_
f
0
_
e
(
_
f
2
dx)
_
2
dx + 2e
(2
_
f
2
dx)
C
1
_
_
_
_
_
_
_
(3.216)
2. Segunda especie clase C
En este caso la ode tiene la siguiente pinta:
(g
1
y + g
0
) y
= f
3
y
3
+ f
2
y
2
+ f
1
y + f
0
(3.217)
La forma de atacar este tipo de odes es la de hacer un cambio de variables
_
x = t, y =
1
u(t)g
1
(t)

g
0
(t)
g
1
(t)
_
(3.218)
qu etrnasforma la ode original en una ode tipo Abel pero de primera clase y se intenta estudiar su tipologa.
La expresin ms general de una ode tipo Chini es:
y
= f
n
y
n
f
1
y + f
0
(3.219)
encontrndose que si n = 2 entonces se combierte en una de tipo Riccati y si n = 3 se combierte en una de tipo
Abel.
Al igual que en el caso de las ode tipo abel si el invariante de Chini
f
n1
n
f
2n+1
0
( f
n
f
0
f
n
f
0
nf
1
f
n
f
0
)
n
n
n
(3.220)
es independiente de x entonces la ode se resuleve de manera directa.
En general y por desgracia este tipo de odes no se deja estudiar por la tcnica de grupos de Lie ya que no se
encuentran soluciones a la edp con la que se obtienen las simetras. No obstante veremos algunos ejemplillos.
3.10.2 Ejemplos
Ejemplo 3.10.1 Solve the following Abel ode
y
=
1 y
2
x y
+ 1 (3.221)
Solution. Introducing the following dimensional constants, we make that eq. (3.221) veries the p.d.h.
(xy) y
= y
2
+ bxy + a, (3.222)
where [b] = yx
1
, [a] = y
2
. Therefore the c.v. that suggests the D.A. is the following one:
_
t =
bxy
a
, u(t) =
y
2
a
_
=
_
x =
at
b
ua
, y =
ua
_
, (3.223)
hence eq. (3.221) yields
u
(2 + t u) t = 2u (1 + t u) , (3.224)
u =
t
2
2t + 2 ln(1 + t) C
1
, (3.225)
in the original variables it yields:
b
2
x
2
2a
=
bxy
a
+ ln(1 +
bxy
a
) + C
1
. (3.226)
Once again we emphasize that the particular solution, y = Cx
1
, (in this case invariant solution , see below) has
been obtained from the relationship
bxy
a
.
Alternatively we can try the following c.v. Writing the original Abel ode, eq.(3.221) in the following form:
axyy
=
y
2
a
+ xy + b, (3.227)
where [a] = XY
1
, [b] = XY, therefore we nd the next c.v.
_
t =
xy
b
, u(t) =
x
2
ab
_
=
_
x =
abu, y =
bt
abu
_
, (3.228)
that brings us to rewrite eq. (3.227) as follows:
u
=
2t
1 + t
= u = 2t 2 ln(1 + t) + C
1
, (3.229)
and hence in the original variables (x, y) we get:
x
2
ab
= 2
xy
b
2 ln(1 +
xy
b
) + C
1
, (3.230)
as we already know. We can check that the following c.v. also works well
_
t =
x
2
ab
, u(t) =
xy
b
_
=
_
x =
abt, y =
bu
abt
_
, (3.231)
since in these variables eq. (3.227) is written as
2u
u = 1 + u = t 2u + 2 ln(1 + u) + C
1
= 0, (3.232)
and therefore
x
2
ab
2
xy
b
+ 2 ln(1 +
xy
b
) + C
1
= 0, (3.233)
obtaining the solution.
Applying the Lie method we need to solve the following pde:
x
+ (
y
x
)(
a y
2
x y
+ 1)
y
(
a y
2
x y
+ 1)
2
+
(a y
2
)
x
2
y
+ (
2
x
+
a y
2
xy
2
) = 0, (3.234)
which solution is:
X
1
=
1
x
x
+
y
x
2
y
= I
1
= xy. (3.235)
This symmetry induces the following c.v.:
_
u(t) =
x
2
2
, t = x y
_
, (3.236)
and therefore in this variables eq. (3.227) is written as follows:
u
=
t
a + t
= u(t) = t + a ln(a + t) + C
1
, (3.237)
hence in the original variables we get:
x
2
2
= x y + a ln(a + x y) + C
1
, (3.238)
This is another example of an ode that it is not scale invariant, and nevertheless, we have been able to reduce it
to a quadrature.
Ejemplo 3.10.2 Solve the following Abel ode.
x (y + 4) y
= y
2
+ 2y + 2x. (3.239)
Solution. Following our pedestrian method, we begin by introducing dimensional constants and rewriting the
odes as follows
x (y + 4a) y
= y
2
+ 2ay + 2bx, (3.240)
where [a] = y, and [b] = x
1
y
2
, therefore D.A. suggest us the following c.v. (note that y =

x, is a particular
solution of (3.239))
_
t =
y
2
bx
, u(t) =
y
a
_
=
_
x =
u
2
a
2
ct
, y = ua
_
, (3.241)
in such a way that eq. (3.239) yields
t (t + 4) u
= t(u + 2) + 2u, (3.242)

which is a linear ode and its solution is:
u = t + C
1
_
t (t + 4), (3.243)
hence
y
a
=
y
2
bx
+ C
1
y
2
bx
_
y
2
bx
+ 4
_
, (3.244)
and simplifying, it is obtained the solution:
bx
a
= y + C
1
_
(y
2
+ 4bx). (3.245)
Another particular solution could be found from the following relationship
y
2
bx
_
y
a
_
1
i.e. y = Cx. But this
particular solution bring us to the following c.v.
_
t =
ay
bx
, u(t) =
y
a
_
=
_
x =
ua
2
ct
, y = ua
_
, (3.246)
which transforms eq. (3.239) into a Bernoulli ode,
t
2
(u + 4) u
= (ut + 2t + 2)
_
u
t u
_
, (3.247)
but this situation is not desirable since it is always more difcult to solve a Bernoulli ode than a linear ode.
Applying the Lie method, following the standard procedure, we need to solve the following pde
x
+
(
y
x
) (y
2
+ 2 y + 2 x)
x (y + 4)

y
(y
2
+ 2 y + 2 x)
2
x
2
(y + 4)
2

(
2
x (y + 4)

y
2
+ 2 y + 2 x
x
2
(y + 4)
) (
2 y + 2
x (y + 4)

y
2
+ 2 y + 2 x
x (y + 4)
2
) = 0, (3.248)
and which solutions are:
X
1
=
_
(2 y + 4 x) (y x)
y + 4
_
y
, X
2
=
_
(y
2
+ 4 x) (y x)
x (y + 4)
_
y
, X
3
=
_
4x + x
2
_
x
+ x (y + 4)
y
(3.249)
where their correspond invariants are:
I
1
= I
2
= x, I
3
=
y + 4
x + 4
(3.250)
For example the c.v. that induces X
3
is:
_
t =
y + 4
4 + x
, u(t) =
1
4
ln(x)
1
4
ln(4 + x)
_
, (3.251)
hence:
_
y =
4 (t 1 + e
(4 u(t))
)
1 + e
(4 u(t))
, x =
4 e
(4 u(t))
1 + e
(4 u(t))
_
, (3.252)
nding that:
u
=
t
2(2t
2
3t + 1)
, (3.253)
which solution is:
u(t) =
1
2
ln(t 1)
1
4
ln(2 t 1) + C
1
, (3.254)
therefore, in the original variables it yields:
1
4
ln(x)
1
4
ln(4 + x) =
1
2
ln(
y + 4
4 + x
1)
1
4
ln(
2 (y + 4)
4 + x
1) + C
1
, (3.255)
as we already know.
Ejemplo 3.10.3 Solve the Abel ode.
y
= Cx
3
y
3
+ Bxy
2
A
y
x
, (3.256)
Solution. If we rewrite eq. (3.256) introducing the following dimensional constants,
y
= a
2
Cx
3
y
3
+ aBxy
2
A
y
x
, (3.257)
where A, B, C R, and [a] = X
2
Y
1
. As we can see this ode is scale invariant since we have needed introduce
only one constant. D.A. suggests us the following c.v.
_
t = x, u(t) = ax
2
y
_
=
_
x = t, y =
u
at
2
_
, (3.258)
in such a way that eq. (3.256) is written in the following form:
tu
= u
_
u
2
+ u + 1
_
, (3.259)
ln t +
1
2
ln
_
u
2
+ u + 1
_
+
3
3
arctan
__
3
2
u +
1
3
_
3
_
ln u + C
1
= 0, (3.260)
therefore in the original variables it yields:
ln x +
1
2
ln
_
_
ax
2
y
_
2
+ ax
2
y + 1
_
+
3
3
arctan
__
3
2
_
ax
2
y
_
+
1
3
_
3
_
ln
_
ax
2
y
_
+ C
1
= 0. (3.261)
In second place, we study eq. (3.256)
y
= Cx
3
y
3
+ Bxy
2
A
y
x
, (3.262)
with respect to the dimensional base B = {T} . This ode veries the principle of dimensional homogeneity with
respect to this dimensional base. Note that [y] =
_
1
H
_
= T
2
, and [x] = [H] = T
1
hence [y
] = T
3
. Therefore
rewriting the equation in a dimensionless way we nd that y x
2
But if we study this equation with respect to the dimensional base B = {X, Y} , we need to introduce new
dimensional constants that make the equation verify the principle of dimensional homogeneity
y
= Cx
3
y
3
+ Bxy
2
A
y
x
(3.263)
where
_
1/2
_
= [] = X
2
Y
1
, hence
y x
X 0 2 1
Y 1 1 0
= y

x
2
, (3.264)
As we can see we have obtained the same solution than in the case of the invariant solution. This is because the
invariant solution that induces a scaling symmetry is the same as the obtained one through the Pi theorem.
This ode admits the following symmetry (scale-invariant)
X = x
x
2y
y
, = I = x
2
y (3.265)
which is a scaling symmetry and it induces the following change of variables,
r = x
2
y, s(r) = ln(x), = x = e
s(r)
, y =
r
e
2s(r)
, (3.266)
which brings us to obtain the next ode in quadratures
s
=
1
r (Cr
2
+ Br + 2 A)
, (3.267)
and which solution is:
s(r) =
lnr
A 2
+
1
2
ln
_
Cr
2
+ Br + 2 A
_
A 2

Barctanh
_
2Cr+B
B
2
+4C(A2)
_
(A 2)
_
B
2
+ 4C(A 2)
+ C
1
, (3.268)
and hence in the original variables (x, y):
ln x =
ln
_
x
2
y
_
A 2
+
1
2
ln
_
Cx
4
y
2
+ Bx
2
y + 2 A
_
A 2

Barctanh
_
2Cx
2
y+B
B
2
+4C(A2)
_
(A 2)
_
B
2
+ 4C(A 2)
+ C
1
, (3.269)
which is the most general solution for this ode.
3.10.3 Pathological cases.
In this section we will present two examples of odes that do not admit symmetries (Lie point symmetries).
Nevertheless in the rst of them D.A. helps us to obtain a simple c.v. that will bring us to obtain a simpler ode
(through a particular solution). In the second case we will show that unfortunately sometimes one nds odes
that at this time have no solution, or at least we do not know how to solve them.
Ejemplo 3.10.4 Solve the Abel ode
_
x
2
y + x
5
x
_
y
= xy
2
_
x
4
+ 1
_
y. (3.270)
Solution. Following our pedestrian method we beginn by introducing dimensional constants
ax
2
yy
+ bx
5
y
cxy
= dxy
2
ex
4
y + f y (3.271)
where
[a] = [d] = x
2
y
1
, [b] = [e] = x
5
, [c] = [ f ] = x
1
, (3.272)
therefore
_
c
5
=
_
f
5
= [b] = [e] , having only two dimensional constants. Since y = x

2
is not a particular
solution of eq.(3.270) then we look for a combination between the monomias nding in this way that
1
=
1
cx
,
2
=
axy
c
(3.273)
where y = x
1
is a particular solution of (3.270). The c.v. that induces the D.A. is the following one:
(t = xy, u(t) = x) =
_
x = u, y =
t
u
_
, (3.274)
hence using these new variables eq. (3.270) is written now as:
2u
t (t + 1) + u
_
1 t u
4
_
= 0, (3.275)
which is a Bernoulli ode and its solution is
u =
_
t
_
2t 2 ln(t 1) + C
1
_
2t 2 ln(t 1) + C
1
, (3.276)
undoing the c.v. we nd that the solution to eq. (3.270) is:
x =
_
xy
_
2xy 2 ln (xy 1) + C
1
_
2xy 2 ln (xy 1) + C
1
. (3.277)
Applying the Lie method we see that the pde to solve is:
x
+ (
y
x
)
y
_
xy x
4
1
_
x (xy + x
4
1)

y
y
2
_
xy x
4
1
_
2
x (xy + x
4
1)
2

_
y
_
y 4x
3
_
x (xy + x
4
1)

y
_
xy x
4
1
_
x
2
(xy + x
4
1)

y
_
xy x
4
1
_ _
y + 4x
3
_
x (xy + x
4
1)
2
_
_
_
xy x
4
1
_
x (xy + x
4
1)
+
y
(xy + x
4
1)

y
_
xy x
4
1
_
(xy + x
4
1)
2
_
= 0. (3.278)
but in this case we have not found any solution, i.e. eq. (3.270) does not admit symmetries. Nevertheless, one
always may try to nd, as if by magic, any c.v. that brings to nd a simpler ode.
In this example we would like to show that there are some odes which are very intractable. Our pedestrian
method does not work in this case, we have not been able to nd any particular solution. The Lie method does
not work, i.e. this ode does not admit any symmetry and the theoretical methods do not work either. At this
time we do not know how to solve it.
Ejemplo 3.10.5 Try to solve the Abel ode
y
= Ax
2
y
3
By
2
y
x
, (3.279)
where A and B are dimensional constant.
Solution. In this occasion we already have the dimensional constants
[A] = x
3
y
2
, [B] = y
1
x
1
, (3.280)
in such a way that all the terms of the equation has dimensions of
y
x
. We check if one of the dimensional relation-
ship induces any particular solution nding that this is not the case. Therefore we go next to look for any trivial
combination between them but we are not able to nd any particular solution. Nevertheless we try to obtain
any result with the following c.v.:
_
t =
B
2
Ax
, u(t) =
1
Byx
_
=
_
x =
B
2
At
, y =
At
B
3
u(t)
_
(3.281)
which bring us to obtain the next ode:
t
2
u
u + tu + 1 = 0, (3.282)
but unfortunately we have not advanced.
Another try is the following one:
_
t =
1
Byx
, u(t) =
B
2
Ax
_
=
_
x =
B
2
Au(t)
, y =
Au(t)
B
3
t
_
(3.283)
and hence:
tu
2
+ u
(1 + ut) = 0, (3.284)
but as we supposed these attemps do not simplify our ode.
Now we change the strategy and we are going to suppose that the ode has dimension of y. In this case we need
to introduce the following dimensional constants in order to make eq. (3.279) verify the p.d.h.
ay
= bx
2
y
3
cy
2
a
y
x
, (3.285)
where [a] = x, [b] = y
2
x
2
and [c] = y
1
, which bring us to the following c.v.
_
t = x, u(t) =
1
y
_
=
_
x = t, y =
1
u(t)
_
(3.286)
therefore we obtain this new ode
u
ut 2u
2
t
5
t
2
u = 0. (3.287)
but as in the above tactic we have not advanced. This c.v. is precisely the suggested one by the theoretic methods.
Following the Lie method we have to solve the pde:
x
+ (
y
x
)
_
Ax
2
y
3
By
2
y
x
_
y
_
Ax
2
y
3
By
2
y
x
_
2
_
1
x
2
y 2Axy
3
_
_
2By
1
x
3Ax
2
y
2
_
= 0, (3.288)
but this ode does not admit any symmetry.
To end, we would like to show that the theoretical method does not work either. For this purpose we follow step
by step the method beginning with a generic Abel ode of rst order written as follows:
y
= f
3
y
3
+ f
2
y
2
+ f
1
y + f
0
, (3.289)
where in our case:
f
3
= x
2
, f
2
= 1, f
1
=
1
x
, f
0
= 0. (3.290)
The c.v. suggested by the theoretical method is the following one
_
t = x, u(t) =
1
y
_
=
_
x = t, y =
1
u(t)
_
, (3.291)
(note that this c.v. is the same than the suggested one by D.A. (3.286)). In this way our ode is now rewritten as:
u
u = h
2
u
2
+ h
1
u + h
0
, (3.292)
where
h
2
= f
1
, h
1
= f
2
, h
0
= f
3
, (3.293)
nding therefore that in our case we have:
u
u =
u
2
t
+ u + t
2
, (3.294)
which is an Abel of second order (this ode has no solution). To try to nd a solution of this ode we make the
following c.v.
(r = t, s(r) = u(t)E) =
_
t = r, u(t) =
s(r)
E
_
, (3.295)
obtaining this new ode:
ss
= F
1
s + F
0
, (3.296)
where
E = exp(
_
h
2
), F
1
= h
1
E, F
0
= h
0
E
2
, (3.297)
hence
ss
=
s
r
+ 1, (3.298)
but unfortunately we do not know how to solve this apparently simple ode.
As we have seen this ode seems very pathological since none of the followed tactics have helped us to obtain
any solution.
3.11 Ecuaciones de DAlambert-Lagrange y de Clairaut
La ecuacin de DAlambert-Lagrange Tiene la forma
y + x f
_
y
_
+
_
y
_
= 0, (3.299)
mientras que la de Clairaut Tiene la forma
y y
x +
_
y
_
= 0. (3.300)
En ambos casos la solucin se obtiene efectuando el siguiente cambio de variable
y
= p.
En el caso de la ecuacin de DAlambert-Lagrange se observa que se transforma en una ecuacin dimensional-
mente homognea, si no se discrimina espacialmente, mediante su simple derivacin respecto a x. Si adems
hacemos z =
dy
dx
, la (3.299) y la que resulta de derivar se escribirn
y + x f (z) + (z) = 0,
z + f (z) +
_
x f
(z) +
(z)
dz
dx
= 0, (3.301)
es decir:
dx
dz
+
f
(z)
z + f (z)
x =

(z)
z + f (z)
, (3.302)
que es una ecuacin lineal de primer orden en x, que sabemos integrar. Sea x = F (z, C) su solucin. Bastar
eliminar z entre ella y (3.301) para encontrar la integral general (x, y, C) de la ecuacin dada.
La ecuacin de Clairaut resulta, por lo tanto un caso particular de la ecuacin de DAlambert-Lagrange, que
venimos de estudiar, en el que
f
_
dy
dx
_
=
dy
dx
, (3.303)
3.11. ECUACIONES DE DALAMBERT-LAGRANGE Y DE CLAIRAUT 77
siguiendo el camino del apartado anterior obtendremos para (3.300) y su derivada respecto x
y xz + (z) = 0,
_
x +
(z)
dz
dx
= 0. (3.304)
Esta ltima ecuacin resultar satisfecha si
dz
dx
= 0 , o sea z = C
1
, que lleva a
y = C
1
x + C
2
, (3.305)
que indica que la solucin de (3.300) corresponde a una familia de rectas. Pero tambin se satisface si
(z) = x.
Entonces de la eliminacin de z entre esta y (3.300,3.304) resultar una curva f (x, y) = 0 que puede demostrarse
que es la envolvente de la familia de rectas dada por (3.305).
Veamos dos ejemplos.
3.11.3 Ejemplos
Ejemplo 3.11.1 Resolver la ode
y = x y
+
m
y
.
Solucin. Veamos que en este caso la edp a resolveres:
x
+
1
2
(
y
x
) (y +
_
y
2
4 x m)
x

1
4
y
(y +
_
y
2
4 x m)
2
x
2

(
y +
_
y
2
4 x m
2 x
2

m
x
_
y
2
4 x m
)
1
2
(1 +
y
y
2
4 x m
)
x
= 0
las soluciones son:
[ = y, = 2 m], [ = 0, =
_
y
2
4 x m],
[ = 0, =
4 x my
2
_
y
2
4 x my
x
],
[ = 0, = 2 x m
y
2
2
+
_
y
2
4 x my
2
],
[ = 2 x, = y], [ = yx, = 2xm + y
2
],
[ = 2 x
2
m + y
2
x, = 3yxm + y
3
]
la simetra [ = y, = 2 m] indice el siguiente c.v. (variables cannicas)
_
t = y
2
4 x m, u(t) =
y
2 m
_
por lo tanto
_
x =
1
4
t + 4 u(t)
2
m
2
m
, y = 2u(t)m
_
u
t
=
1
4
t m
= u(t) =
1
2
t
m
+ C
1
la solucin es:
y
2 m
=
1
2
_
y
2
4 x m
m
+ C
1
simplicando se obtiene
y
2
4 x m = 0, e y = xC
1
+
m
C
1
.
Observamos que existe una simetra de escala [ = 2 x, = y] que induce la siguiente solucin invariante:
dx
2x
=
dy
y
= y
2
= ax
donde a es cierta constante numrica. Esta es precisamente la solucin que obtendramos utilizando el teorema
Pi. Como se observa esta solucin es precisamente la envolvente de la familia de rectas dada por: y = xC
1
+
m
C
1
.
El cv que induce el AD es el siguiente:
t = x, u(t) = y
2
x
1
, =
ut =
1
2
t (u + tu
ut
+
2
ut
u + tu
cuya solucin es:

ln t ln
_
u 2 +
_
u
2
4u
_
C
1
= 0,
y por lo tanto obtenderemos
ln x ln
_
y
2
x
2 +
_
y
4
x
2
4
y
2
x
_
C
1
= 0,
Ejemplo 3.11.2 Resolver la ode
y = xy
2y
2
Solucin. Veamos que
x
+ (
y
x
) (
x
4
+
_
x
2
8 y
4
)
y
(
x
4
+
_
x
2
8 y
4
)
2
(
1
4
+
x
4
_
x
2
8 y
) +

_
x
2
8 y
= 0
la solucin es
[ = 1, =
x
4
], [ = 0, =
_
x
2
8 y],
[ = 0, =
x
2
4
+
x
_
x
2
8 y
4
2 y],
[ = 0, = 2 y x y
_
x
2
8 y +
x
3
4
+
x
2
_
x
2
8 y
4
],
[ = x, = 2 y], [ = 4 y + x
2
, = y x],
[ = x (6 y + x
2
), = y (4 y + x
2
)]
Vemos que [ = 1, =
x
4
] indice el sigiente c.v.
_
u(t) = x, t = y
x
2
8
_
3.12. CONCLUSIONS AND DISCUSSION. 79
por lo tanto
y = t +
1
8
u(t)
2
, x = u(t)
u
t
=
2
2 t
= u(t) = 2
2
_
(t) + C
1
de esta forma llegamos a la siguiente solucin en las variables originales:
y =
x
2
8
, e y = C
1
x 2C
2
1
observamos que la primera de las soluciones es precisamente la que obtenemos como solucin invariante in-
ducida por [ = x, = 2 y].
El cv que iduce el AD es el siguiente:
t = x, u(t) = y
1
x
2
, =
t
2
u
= t
_
2t
u

t
2
u
u
2
_
2
_
2t
u

t
2
u
u
2
_
2
cuya solucin es:
ln t
1
2
ln
_
4 + u +
_
u
2
8u
_
1
2
ln u C
1
= 0,
y por lo tanto obtenderemos
ln x
1
2
ln
_
x
2
y
2 +
x
4
y
2
8
x
2
y
_
1
2
ln
x
2
y
C
1
= 0,
3.12 Conclusions and discussion.
We have seen howwriting the odes in such a way that they verify the pdh i.e. introducing dimensional constants,
we can obtain in a trivial way c.v. that bring us to obtain simpler ode than the original and therefore their
integration is immediate. Furthermore, we have tried to show that these c.v. are not obtained as if by magic
but that they correspond to invariant solutions or to particular solutions and therefore they are generated by the
symmetries that admit the ode.
Nevertheless, the D.A. has strong limitations. For example D.A. is unable (at least at this time we do not know
how to do it) to solve the following simple linear ode
y
=
_
x
3
+
1
x
+ 3
_
y +
_
3x
2
1
x
2
_
, (3.306)
for this reason one must not put all his condence in this tactic. This is one of the greater inconvenience that
presents the proposed method. But as we have noticed in the introduction, we think that our pedestrian method
continues having validity at least when one is studying ode derived from engineering problems or physical
problems etc... where, as it is supposed, such odes must verify the pdh in such a way that for example the ode
y
= x + 1, lacks of any sense (physical sense, since we cannot add a number to a physical quantity).
Nevertheless, and in spite of the limitations that we have not avoided to show, we continued believing in the
kindness (goodness) of the method and that it can be applied to obtain solutions (at least particular solutions
and in concrete invariant solutions) to more complicated equations like the following ones:
y
=
y
2
y
ay
2
, [a] = y
1
x
2
= y =
1
x
2
, (3.307)
y
=
a
y
3
, [a] = y
4
x
3
= y = x
3/4
, (3.308)
y
= ayy
, [a] = y
1
x
1
= y = x
1
, (3.309)
y
=
y
x
+ a
3
2
y
2
x
3
, [a] = y
1
x = y = x, (3.310)
y
=
(y
)
2
y
(1 + ay
)
, [a] = y
1
x = y = x, (3.311)
but these are questions that we will approach in a forthcoming paper.
Bibliography
[1] A.E. Ruak. Jour. of the Mitchell Socety. August (1935), 127-133.
[2] G. Birhoff. Electrical Egineering. December (1948), 1185-1188.
G. Birhoff. Hydrodynamics. Princeton (1950)
[3] F.A. Morrison, Jr. Bull Mech. Eng. Edu. 8, (1969), 289-300.
[4] M.J. Moran. Jour. Franklin Inst. 292, (1971), 423-432.
[5] J. Hainzl. Jour. Franklin Inst. 292, (1971), 463-470.
[6] T.Y. Na and A.G. Hansen. Jour. Franklin Inst. 292, (1971), 471-489.
[7] G.I. Barenblatt and Y.B. Zeldovich. (1972) 285-312
[8] K. Kurt. Dimensional Analysis and Group Theory in astrophysics. Pergamon Press Oxford (1972).
[9] R. Seshadri and T.Y. Na. Group Invariance in Enginenering Boundary Value Problems. Springer-Verlang. N.Y. (1985) .
[10] G.I. Bararenblatt. Scaling, Self-Similarity and Intermediate Asymptotics. CUP (1996)
[11] M. Castas and C. Daz. J. Non-Equilib. Thermodyn. 14, (1989) 315-329.
M. Castas and C. Daz. Il Nuovo Cimento D 13, (1991), 1061-1069.
[12] J.A. Belinchn and P. Dvila. Class. Quantum Grav. 17,(2000), 3183-98.
J.A. Belinchn . gr-qc/0404028.
[13] M. Castas and P. Dvila. GIAD. Lanzarote 1999.
[14] E.J. Wilczynski. Trans. Am. Math Soc 23, (1900).
[15] A. Cohen. An Introduction to the Lie Theory of one-parameter groups. Heath, N.Y. (1911).
[16] A.J.A. Morgan. Quart. Jour. Math. Oxford (2), 3 (1952) 250-59
[17] L. P. Eisenhart. Continuous Groups of Transformations. Dover 2003.
[18] L.V. Ovsiannikov. Group Analysis of differential Eqautions. Academic. Press, N.Y. (1982).
[19] N.H. Ibragimov. Elementary Lie Group Analysis and Ordinary Differential Equations. Chicherter. John Wiley & Sons (1999).
N.H. Ibragimov. Introduction to Modern Group Analysis. Ufa. (2000).
[20] G.W.Bluman and J.D. Cole. Similarity Methods for Differential Equations. Springer-Verlang (1974).
G.W.Bluman and S. Kumei. Symmetries and Differential Eqautions. Springer-Verlang (1989).
G.W.Bluman and S.C. Anco. Symmetry and Integration Methods for Differential Equations. Springer-Verlang. (2002).
[21] P.E. Hydon. Symmetry Methods for Differential Equations. CUP, (2000)
81
82 BIBLIOGRAPHY
[22] B.J. Cantwell. Introduction to Symmetry Analysis". CUP. (2002).
[23] G. Baumann. Symmetry Analysis of Differential Equations with Mathematica. Springer Telos. (2000).
[24] H.Stephani. Differential Equations. Their Solution Using Symmetries. CUP (1989).
[25] G. Emanuel. Solution of Ordinary Differential Equations by Continuos Groups. Chapma&Hall/CRC (2001).
[26] J.M. Hill. Differential Equations and Group Methods forScintists and Engineers. CRC (1992).
[27] P.J. Olver. Applications of Lie Groups to Differential Equations. 2nd ed. Springer-Verlang (1993).
[28] E. Kamke. Differentialgleichungen: Losungsmethoden und Losungen, Akademische Verlagsgeselischaft Leipzig (1959).
[29] G.M. Murphy. Ordinary Differential Equations and their Solutions. Princeton: Van Nostrand, (1960).
[30] H.T. Davis Introduction to Nonlinear Differential and Integral Equation Dover N.Y. (1962).
[31] D. Zwillinger. Handbook of Differential Equations. 2nd edition. AcademicPress, (1992).
[32] A.D. Polyanin and V.F. Zaitsev. Handbook of Exact Solution for Ordinary Differential Equations. CRC Press Boca Raton FL (1995).
[33] E.S. Cheb-Terrab and A.D. Roche. : European Journal of Applied Mathematics (2003), vol. 14, pp. 217-229. math.GM/0002059.
E.S. Cheb-Terrab and A.D. Roche. Computer Physics Communications. math-ph/0001037
E.S. Cheb-Terrab and T. Kolokolnikov. math-phys/0007023.
[34] F. Schwarz. Studies in Applied Mathematics 100, (1998), 269-294.
[35] V.M. Boyko. nlin.SI/0404020
[36] M.K. Mak, H.W. Chan and T. Harko. Computers & Mathematics, 41 (2001) 1395-1401.
M.K. Mak and T. Harko. Computers & Mathematics, 43 (2002) 91-94.
T. Harko and M.K. Mak. Computers & Mathematics, 46 (2003) 849-853.
[37] M.L. Green and P. A. Grifths. Houston Journal of Mathematics 28, (2002) 329-351.
Chapter 4
Ecuaciones de segundo orden.
4.1 Introduccin
Presentamos un breve resumen del mtodo de Lie aplicado a ecuaciones de segundo orden. Explicamos como
reducir el orden, integrarlas o linealizarlas.
Reduccin del orden.
Si una ecuacin de segundo o mayor orden admite al menos una simetra entonces se puede reducir su orden en
uno. Para una ODE
y
= f
_
x, y, y
_
, (4.1)
admite una simetra
X = (x, y)
x
+ (x, y)
y
, (4.2)
puede ser integrada una vez introduciendo variables cannicas o usando el mtodo de la diferenciacin invari-
ante.
Variables cannicas.
La introduccin de variables cannicas, elimina la dependencia explcita de (4.1) de una de las variables x o y.
Por lo tanto, una ODE de segundo orden con una simetra conocida puede ser convetida en una de las siguientes
formas
y
= f
_
y, y
_
, o y
= f
_
x, y
_
, (4.3)
inmediatamente reducible a una ODE de primer orden mediante mtodos elementales.
Diferenciacin invariante.
El mtodo basado en el siguiente teorema proporciona una generalizacin y una formulacin covariante (i.e.
independiente de cambios de variable de x o y) de mtodos ad hoc para bajar el orden de una ecuacin.
Teorema 4.1.1 La ODE (4.1) que admite un grupo uni-paramtrico, G, generado por la simetra, X, dada por la eq. (4.2)
puede ser escrita en la forma de una ODE de primer orden
dv
du
= F (u, v) , (4.4)
con una funcin arbitraria F (u, v) . Aqu, u = u (x, y) denota un invariante del grupo G y v = v (x, y, y
) su diferencial
invariante de primer orden.
83
84 CHAPTER 4. ECUACIONES DE SEGUNDO ORDEN.
Integracin por medio de dos simetras.
De acuardo con la clasicacin de grupos de Lie de ODEs, tenemos que el lgebra de Lie maximal L
r
admitida
por una ODE (4.1) slo puede tener dimensiones r = 1, 2, 3 u 8. El caso r = 8 indica que (4.1) es lineal o
linealizable. Si (4.1) admite L
r
con r 2, podemos escoger una sublgebra 2dimensional L
2
L
r
. Por lo tanto
asumiremos que (4.1) admite una L
2
.
Teorema 4.1.2 Una ODE (4.1) admite una lgebra de Lie L
2
, con una base X
1
, X
2
, tal que el lgebra L
1
generado por X
1
es un ideal en L
2
. Entonces la ecuacin resultante de reducir la ODE (4.1) mediante X
1
admite el lgebra cociente L
2
/L
1
.
Es decir, dada la base X
1
, X
2
, entonces calculamos [X
1
, X
2
] = cX
1
. Entonces X
1
genera el ideal L
1
mientras que el
lgebra cociente L
2
/L
1
puede seridenticada con el lgebra que genera X
2
. La ODE de primer orden resultante
de reducir el orden mediante X
1
toma la forma dada por (4.4) entonces esta ODE admite X
2
pero escrito en
trmino de las variables u, v.
Forma cannica de L
2
.
La forma cannica de L
2
nos suministra el mtodo ms simple de integracin de una ODE pues las reduce a slo
4 tipos.
Teorema 4.1.3 Toda L
2
puede reducirse, eligiendo una base apropiada,
X
1
=
1
x
+
1
y
, X
2
=
2
x
+
2
y
,
a uno de los siguientes tipos:
I [X
1
, X
2
] = 0
1
2
= 0,
II [X
1
, X
2
] = 0
1
2
= 0,
III [X
1
, X
2
] = X
1

1
2
= 0,
IV [X
1
, X
2
] = X
1

1
2
= 0.
Este teorema nos permite obtener la siguiente tabla:
Estructura de L
2
Forma cannica de L
2
ODE
I [X
1
, X
2
] = 0,
1
2
= 0, X
1
=
x
, X
2
=
y
, y
= f (y
) ,
II [X
1
, X
2
] = 0,
1
2
= 0, X
1
=
y
, X
2
= x
y
, y
= f (x) ,
III [X
1
, X
2
] = X
1
,
1
2
= 0, X
1
=
y
, X
2
= x
x
+ y
y
, y
=
1
x
f (y
) ,
IV [X
1
, X
2
] = X
1
,
1
2
= 0, X
1
=
y
, X
2
= y
y
, y
= f (x) y
.
La receta
Los paso a seguir son por lo tanto los siguientes:
1. Calcular el lgebra de Lie que admite L
r
.
2. Si r 2, entonces calcular L
2
L
r
. Si r < 2 entonces no se puede aplicar el mtodo.
3. Determinar el tipo de L
2
(mirar la tabla anterior).
4. Encontrar el cambio de variable (c.v.) que tranforma L
2
en su forma cannica.
5. integrar la ecuacin resultante en las nuevas variables.
4.2. DISTINTOS TIPOS DE ECUACIONES. 85
4.2 Distintos tipos de ecuaciones.
4.2.1 Lineales con coeciente constantes.
Sea dada la ecuacin diferencial
a
0
y
n
+ a
1
y
n1
+ ......... + a
n
y = 0 (4.5)
donde las a
i
son constantes reales.
Consideramos la ecuacin caracterstica
a
0
n
+ a
1
n1
+ ......... + a
n
= 0 (4.6)
supongamos que las (
i
) son las races de la ecuacin, entre las cuales puede haber mltiples. Se pueden pre-
sentar los siguientes casos.
1. Las races son reales y distintas i.e.
1
, .....,
n
son reales y distintas
En este caso el sistema fundamental de soluciones tiene la forma
e
1
x
, .........., e
n
x
(4.7)
y la solucin general a la ecuacin homognea es :
y
g
= C
1
e
1
x
+ .......... + C
n
e
n
x
(4.8)
2. Las raices son reales pero algunas de ellas son mltiples. Sea por ejemplo
1
=
2
= ... =
k
=

, (4.9)
de modo que

es una raiz kmltiple , mientras que todas las dems (n k) raices son distintas. En este
caso el sistema fundamental de soluciones tiene la forma
e
x
, xe
x
, x
2
e
x
, ........, x
k1
e
x
, e
k+1
x
, ........, e
n
x
, (4.10)
y la solucin general es :
y
g
= C
1
e
x
+ C
2
xe
x
+ C
3
x
2
e
x
+ ... + C
k
x
k1
e
x
+ C
k+1
e
k+1
x
+ ... + C
n
e
n
x
. (4.11)
3. Supongamos que algunas de las raices son imaginarias. Para simplicar imaginemos que la situacin es la
siguiente :
1
= + i,
2
= i,
3
= + i,
4
= i, (4.12)
y las dems raices son todas reales. En este caso el sistema fundamental de ecuaciones es :
e
x
cos , e
x
sen , e
x
cos , e
x
sen , e
5
x
, ......., e
n
x
(4.13)
y la solucin general es :
y
g
= C
1
e
x
cos + C
2
e
x
sen + C
3
e
x
cos + C
4
e
x
sen + C
5
e
5
x
+ .. + C
n
e
n
x
(4.14)
4. Si
1
= + i
es una raiz k-mltiple
k <
_
n
2
_
entonces,
2
= i
ser tambien una raiz k-mltiple y el sistema fundamental de soluciones ser de la forma :
e
x
cos , e
x
sen , xe
x
cos , xe
x
sen , ......, x
k1
e
x
cos , x
k1
e
x
sen ,
e
2k+1
x
, ......., e
n
x
(4.15)
por consiguiente la solucin general es de la forma
y
g
= C
1
e
x
cos + C
2
e
x
sen + C
3
xe
x
cos + C
4
xe
x
sen + .......+
C
2k1
x
k1
e
x
cos + C
2k
x
k1
e
x
sen + C
2k+1
e
2k+1
x
+ ............. + C
n
e
n
x
(4.16)
Ejemplos.
Vamos a estudiar unos ejemplos para el caso en el que la ecuacin no es homognea i.e. tiene la forma
a
0
y
n
+ a
1
y
n1
+ ......... + a
n
y = f (x), (4.17)
donde la f (x) es cierta funcin de x.
Atendiendo a la anterior clasicacin i.e. adependiendo de los autovalores y sus respectivas multiplicidades nos
podemos encontrar las siguiente tabla
c f (x) raices de la ecuacin y
p
(x)
1 P
m
(x)
1. 0 / ()
2. 0 (), (0) = s.
P
m
(x)
x
s

P
m
(x)
2 P
m
(x)e
x
1. / ()
2. (), () = s.
P
m
(x)e
x
x
s

P
m
(x)e
x
3 P
m
(x) cos x + Q
m
(x) sin x
1. i / ()
2. i (), (i) = s.
P
m
(x) cos x +

Q
m
(x) sin x
x
s
_

P
m
(x) cos x +

Q
m
(x) sin x
_
4 e
x
[P
m
(x) cos x + Q
m
(x) sin x]
1. + i / ()
2. + i (), /
( + i) = s.
e
x
_

P
m
(x) cos x +

Q
m
(x) sin x
_
e
x
x
s
_

P
m
(x) cos x +

Q
m
(x) sin x
_
Con esta tabla en mente pasaremos a estudiar algunos ejemplillos.
y
+ y
y = x
2
+ x (4.18)
Solution. Las raices de la ecuacin homognea son (1, i, i) , por lo que la solucin de la homognera ser de la
forma
y
h
= C
1
e
x
+ C
2
sin x + C
3
cos x. (4.19)
Al ser f (x) = P
m
(x) = x
2
+ x i.e. un polinomio y 0 / (), entonces la forma de la solucin particular ser
y
p
= A
1
x
2
+ A
2
x + A
3
,
esta ecuacin debe debe vericar la ecuacin original as que
y
p
y
p
+ y
p
y
p
= x
2
+ x
(2A
1
) + (2A
1
x + A
2
)
_
A
1
x
2
+ A
2
x + A
3
_
= x
2
+ x,
por lo tanto
A
1
= 1, A
2
= 3, A
3
= 1.
De esta forma la solucin general es:
y = C
1
e
x
+ C
2
sin x + C
3
cos x x
2
3x 1. (4.20)
como queramos demostrar.
y
= 12x
2
+ 6x. (4.21)
Solution. Las raices de la ecuacin homognea son (1, 0, 0) , por lo que la solucin de la homognera ser de la
forma
y
h
= C
3
e
x
+ C
2
x + C
1
. (4.22)
Al ser f (x) = 12x
2
+ 6x i.e. un polinomio y 0 (), (0) = s = 2, entonces la forma de la solucin particular
ser
y
p
= x
s

P
m
(x) = x
2
_
A
1
x
2
+ A
2
x + A
3
_
=
_
A
1
x
4
+ A
2
x
3
+ A
3
x
2
_
,
esta ecuacin debe debe vericar la ecuacin original as que
y
p
y
p
+ y
p
y
p
= 12x
2
+ 6x
por lo tanto
A
1
= 1, A
2
= 5, A
3
= 15.
y = C
1
+ C
2
x + C
3
e
x
x
4
5x
3
15x
2
, (4.23)
y
+ y
= 2x
2
e
x
. (4.24)
Solution. Las raices de la ecuacin homognea son (1, 0) , por lo que la solucin de la homognera ser de la
forma
y
h
= C
1
+ C
2
e
x
. (4.25)
Al ser f (x) = 2x
2
e
x
i.e. un polinomio y = 1 / (), entonces la forma de la solucin particular ser
y
p
= e
x

P
m
(x) = e
x
_
A
1
x
2
+ A
2
x + A
3
_
,
por lo tanto
A
1
= 2, A
2
= 6, A
3
= 7.
y = C
1
+ C
2
e
x
+
_
2x
2
6x + 7
_
e
x
, (4.26)
y
+ 10y
+ 25y = 4e
5x
. (4.27)
Solution. Las raices de la ecuacin homognea son (5, 5) , por lo que la solucin de la homognera ser de
la forma
y
h
= (C
1
+ C
2
x) e
5x
. (4.28)
Al ser f (x) = 4e
5x
i.e. un polinomio y = 5 (), () = s, entonces la forma de la solucin particular
ser
y
p
= e
x

P
m
(x)x
s
= e
5x
(A
1
) x
2
,
por lo tanto
A
1
= 2, .
y = (C
1
+ C
2
x) e
5x
+ 2x
2
e
5x
, (4.29)
y
+ 3y
+ 2y = x sin x. (4.30)
la forma
y
h
= C
1
e
x
+ C
2
e
2x
. (4.31)
Al ser f (x) = x sin x i.e. un polinomio y i / (), entonces la forma de la solucin particular ser
y
p
=

P
m
(x) = e
5x
(A
1
) x
2
,
por lo tanto
A
1
= 2, .
y = (C
1
+ C
2
x) e
5x
+ 2x
2
e
5x
, (4.32)
y
+ 3y
+ 2y = x sin x. (4.33)
la forma
y
h
= C
1
e
x
+ C
2
e
2x
. (4.34)
Al ser f (x) = 4e
5x
i.e. un polinomio y = 5 (), () = s, entonces la forma de la solucin particular
ser
y
p
= e
x

P
m
(x)x
s
= e
5x
(A
1
) x
2
,
por lo tanto
A
1
= 2, .
y = (C
1
+ C
2
x) e
5x
+ 2x
2
e
5x
, (4.35)
4.2.2 Ecuacin de Bessel.
La ecuacin de Bessel tiene la forma
x
2
y
+ xy
+
_
x
2
n
2
_
y = 0 (4.36)
y cuya solucin es:
y = C
1
Bessel J(n, x) + C
2
BesselY(n, x) (4.37)
La ecuacin de Bessel modicada es de la forma:
x
2
y
+ xy
_
x
2
+ n
2
_
y = 0 (4.38)
y cuya solucin es:
y = C
1
Bessel I(n, x) + C
2
BesselK(n, x) (4.39)
Estas ecuaciones no admiten simetras de tal forma que el mtodo de Lie es inoperante.
4.2.3 Ecuacin de Dufng.
La ecuacin de Dufng tiene la forma
y
+ y
+ y
3
= 0 (4.40)
y cuya solucin es:
_
y
2da
_
4a
2
2a
4
+ 4C
1
x C
2
= 0. (4.41)
Esta ecuacin admite una simetra, X
1
=
x
, que nos lleva a obtener el siguiente c.v.
t = y, u = 1/y
, (4.42)
de donde obtenemos,
u
= t
_
1 + t
2
_
u
3
, (4.43)
i.e. una ODE separable y cuya solucin es:
u =
2
_
4t
2
2t
4
+ 4C
1
. (4.44)
4.2.4 Ecuacin de Emden y Emden-Fowler.
Vamos a estudiar tres tipos de ecuaciones tipo Emden. Para ms detalles y ms ecuaciones recomendamos el
libro de A. Polyanin y V. Zaitsev, Exact Solutions for ODEs, CRC 1995, pginas 241-301.
1. La ecuacin de Emden tiene la forma
x
2
y
+ 2xy
+ x
2
y
n
= 0, (4.45)
y cuya solucin es:
y = te
2
1n
_
udt+C
1
, x = e
_
udt+C
1
, (4.46)
donde u(t) es solucin de la ODE (tipo Abel)
u
=
u
2
(n 1)
2
_
n
2
(1 + ut
n
) + 2n (ut + 3 + ut
n
) ut
_
6 + t
n1
_
5
_
. (4.47)
que no admite simetras y que no conocemos su solucin. Vemos adems que la ecuacin (4.45) admite
una nica simetra, X
1
= (1 n)x
x
+ 2y
y
, y por lo tanto debe admitir la solucin particular
y = ax
2/(1n)
, (4.48)
observndose que la eq. (4.45) es DH sii introducimos una constante a, tal que
x
2
y
+ 2xy
+ ax
2
y
n
= 0, [a] = y
1n
x
2
. (4.49)
2. La ecuacin de Emden modicada tiene la forma
y
+ f
1
(x)y
+ y
n
= 0, (4.50)
y cuya solucin es:
y =?, (4.51)
dependiendo de la forma de f (x) la ecuacin podr admitir una u otra simetra. Tambin podemos calcular
el sistema de EDPs para determinar sus simetras e imponer alguna determinada, como por ejemplo la
de escala, de esta forma hayamos las posibles formas que puede tomar f (x) para que la ecuacin sea
integrable. Ntese que si imponemos la simetra de escala entonces f (x) = x
1
, y por lo tanto estaramos
en la situacin anteriror.
3. La ecuacin de Emden-Fowler tiene la forma
x
p
y
+ x
p1
py
+ x
y
n
= 0, (4.52)
i.e.
y
=
p
x
y
ax
p
y
n
, [a] = y
1n
x
p2
.
de esta forma vemos que una solucion particular debe ser:
y = ax
2+p
1n
. (4.53)
Al igual que en la ecuacin de Emden, en este caso, la ecuacin de Emden-Fowler tambin admite una
simetra de escala, tal y como hemos visto, X = (1 n)x
x
+ (2 + p) y
y
. El c.v. obtenido por las
variables cannicas nos lleva a obtener una ODE tipo Abel sin solucin.
Observacin 4.2.1 En las ecuaciones que slo admitan una simetra y sta sea de escala, como en las ecuaciones tipo
Emden y EF, el cv que induce el AD es el mismo que el de grupos de Lie. Si la ODE admite ms simetras y el campo
resultante (el que es un ideal del lgebra) no es de escala, entonces el cv que iduce el AD no funciona, es ms complica
mucho la situacin. Veremos algunos ejemplos ms adelante.
4.2.5 Ecuacin lineal exacta.
La ecuacin lineal exacta tiene la forma
y
= f (x)y
+ f
(x)y, (4.54)
y
y
= f (x)y
+ f
(x)y + g
(x), (4.55)
donde f
= d f /dx, y cuya solucin de (4.55) es:

y =
_
C
2
+
_
_
C
1
+ ge
_
f dx
_
dx
_
e
_
f dx
. (4.56)
4.2.6 Ecuacin no lineal exacta.
Un ejemplo de ecuacin no lineal exacta puede ser el siguiente
y
=
x
y
2
y
+
1
y
, (4.57)
cuya solucin de es:
1
2
ln(
x
2
C
1
y x y
2
x
2
) +
C
1
arctanh
_
C
1
x+2 y
x
4+C
2
1
_
_
4 + C
2
1
+ ln(x) + C
2
= 0. (4.58)
Observamos que podemos rescribir la ecuacin de la siguiente forma
y
=
ax
y
2
y
+
a
y
, [a] = y
2
x
2
, (4.59)
por lo que una solucin particular a dicha ecuacin ser:
y = x. (4.60)
Veamos cuales son las simetrias de la ODE
y y
y
3
= 0,
2 y
y
x +
y, y
y
3
2
x y
y
3
= 0,
2 x 3
y
y
2
+ y
x
x + y + 2
xy
y
3
xx
y
3
= 0,
y 2
x
y
2
+
y
y
2
+ x y
x
+
x x
y
3
= 0,
por lo tanto
= x, = y
el cambio de variables que induce esta simetra es el siguiente
_
u(t) = ln(x), t =
y
x
_
=
_
x = e
u(t)
, y = t e
u(t)
_
por lo tanto la ODE se rescribe como
u
t, t
=
u
t
2
(t
2
+ 1)
t
2
y cuya solucin es
u(t) =
1
2
ln(t
2
1 C
1
t) +
C
1
arctanh(
2 tC
1
4+C
2
1
)
_
4 + C
2
1
,
y por lo tanto en las variables originales obtenemos la solucin antes expuesta.
La reduccin de orden viene dada por el siguiente c.v.
t =
y
x
, u(t) =
x
y + xy
,
por lo tanto la ODE original queda
u
t
=
u(t)
2
(t
2
+ 1)
t
2
viendo que se trata de una ODE separable, cuya solucin es:
u(t) =
t
t
2
1 C
1
t
,
haciendo el c.v. en las variables originales obtenemos
y
=
x
y
+ C
1
cuya solucin es (4.58).
En este ejemplo hemos visto que la ODE bajo estudio slo admite una simetra y que sta, adems es de escala.
Por lo tanto el cv que induce el AD es el mismo que el de GL ie.
t =
y
x
, u(t) =
x
y + xy
.
Tambin podramos emplear una tctica del estilo
_
t = x, u(t) =
y
x
_
y ver que pasa.
4.2.7 La ecuacin lineal.
La ecuacin lineal tiene la forma
y
= p(x)y
+ q(x)y. (4.61)
La condicin de linealizacin nos lleva a:
= Ay + B,
=
_
A
+ pA
_
y
2
+ Cy + D,
donde
A
+ (pA)
qA = 0,
B
+ (pB)
= 2C
,
C
pC
= 2qB
+ q
B,
D
pD
qD = 0,
viendo que la ecuacin para D es la propia ODE, mientras que A satisface la adjunta. Supongamos ahora que la
solucin general es de la forma
y = K
1
y
1
+ K
2
y
2
, K
i
R,
entonces
A = e
_
pdx
(c
4
y
1
+ c
5
y
2
) ,
B = e
_
pdx
_
c
6
y
2
1
+ 2c
7
y
1
y
2
+ c
8
y
2
2
_
,
C = c
1
+ e
_
pdx
_
c
6
y
1
y
1
+ c
7
_
y
1
y
2
+ y
1
y
2
_
+ c
8
y
2
y
2
_
,
D = c
2
y
1
+ c
3
y
2
, ,
por lo que nuestra ODE admite las siguientes simetras:
X
1
= y
y
, X
2
= y
1
y
, X
3
= y
2
y
, X
4
= e
_
pdx
_
yy
1
x
+ y
1
y
2
y
_
,
X
5
= e
_
pdx
_
yy
2
x
+ y
2
y
2
y
_
, X
6
= e
_
pdx
_
y
2
1
x
+ y
1
y
1
y
y
_
,
X
7
= e
_
pdx
_
2y
1
y
2
x
+
_
y
2
y
1
+ y
2
y
1
_
y
_
, X
8
= e
_
pdx
_
y
2
2
x
+ y
2
y
2
y
y
_
.
4.2.8 Ecuacin de Largerstrom.
La ecuacin de Largerstrom tiene la forma
y
=
K
x
y
yy
, (4.62)
i.e.
y
=
_
K
x
+ y
_
y
, [K] = x
0
y
0
, [] = y
1
x
1
,
y cuya solucin es:
y = te
(
_
u(t)dt+C
1
)
, u
= (2 K t) tu
3
+ (3 K t) u
2
, (4.63)
donde se observa que esta ODE es de tipo Abel y no tiene solucin.
La ecuacin (4.62) admite simetra de escala
X = x
x
y
y
,
que induce el c.v.
u =
1
(y
x + y) x
, t = xy,
con el que se obtiene la ODE (4.63).
De igual forma vemos que la ODE (4.62) admite solucin particular
y =
c
x
, c R.
4.2.9 Ecaucin de Lienard.
La ecuacin de Lienard tiene la forma
y
+ f (x)y
+ y = 0 (4.64)
cuya solucin es:
y = e
(
_
u(t)dt+C
1
)
, u
= u
2
f u 1, (4.65)
es una ODE tipo Riccati. La reduccin de orden viene dada por el siguiente c.v.
u =
y
y
, x = t. (4.66)
Veamos un ejemplo concreto.
y
+
1
x
y
+ Ky = 0
vemos que su solucin es:
y = C
1
Bessel J (0, x) + C
2
Bessel(0, x).
4.2.10 La ecuacin de Liouville.
La ecuacin de Lioville tiene la forma
y
+ g(y)y
2
+ f (x)y
= 0, (4.67)
esta ecuacin admite las siguientes simetras
X
1
= e
_
gdy
y
, X
2
= e
_
f dx
x
, (4.68)
y cuya solucin es:
_
y
e
_
g(b)db
db C
1
_
e
_
f (x)dx
dx C
2
= 0. (4.69)
La reduccin de oreden viene dada por el c.v.
u
= f (t)u, t = x, u = e
_
y
g(b)db
. (4.70)
4.2.11 La ecuacin de Painleve.
Las ecuaciones de Painleve se dividen en 6 grupos. Exponemos a continuacin ejemplos de cada una de las
clases.
1. Painleve I
y
= 6y
2
+ x, (4.71)
2. Painleve II
y
= 2y
3
+ xy + a, (4.72)
3. Painleve III
y
=
y
2
y

y
x
+
ay
2
+ b
x
+ cy
3
+
d
y
, (4.73)
4. Painleve IV
y
=
1
2
y
2
y
+
3
2
y
3
+ 4xy
2
+ 2
_
x
2
a
_
y +
b
y
, (4.74)
5. Painleve V
y
=
_
1
2y
+
1
y 1
_
y
2
x
+
(y 1)
2
_
ay +
b
y
_
x
2
+
gy
x
+
dy (y + 1)
y 1
, (4.75)
6. Painleve VI
y
=
_
1
y
+
1
y 1
+
1
y x
_
y
2
_
1
x
+
1
x 1
+
1
y x
_
y
+
+
y (y 1) (y x)
_
a +
bx
y
2
+
g(x1)
(y1)
2
+
dx(x1)
(yx)
2
_
x
2
(x 1)
2
. (4.76)
Todas estas ecuaciones son complicadas de resolver y ninguna de ellas admite simetras por lo que el mtodo de
Lie no funciona.
4.2.12 La ecuacin de Titchmarch.
La ecuacin de Titchmarch tiene la forma
y
+
_
x
2n
_
y = 0 (4.77)
y cuya solucin es:
y = e
_
u(t)dt+C
1
. (4.78)
Esta ODE admite una simetra, X = y
y
, que induce el siguiente c.v.
t = x, u(t) =
y
y
, (4.79)
por lo tanto obtenemos
u
= u
2
+ t
2n
, (4.80)
que es tipo Riccati, pero sin solucin.
4.2.13 La ecuacin de Van der Pol.
La ecuacin de Van der Pol tiene la forma
y
_
1 y
2
_
y
+ y = 0 (4.81)
que admite una simetra
X =
x
, (4.82)
que nos lleva a obtener el siguiente c.v.
_
t = y, u(t) =
1
y
_
, =
_
y = t, x =
_
udt + C
1
_
, (4.83)
por lo que la ODE original se reduce a:
u
= tu
3
+
_
t
2
1
_
u
2
, (4.84)
tipo Abel sin solucin.
4.3 Ejemplos
Ejemplo 4.3.1 Resolver la siguiente ecuacin de segundo orden
y
=
y
2
y
+ y
2
(4.85)
Solucin. Aplicando el mtodo de Lie lo primero que tenemos que planear el es sistema de edps que nos
determinarn las simetras, este sistema de edps es:
y
yy
+
y
= 0,
y
y
+ y
2
yy
2y
2
xy
= 0,
3y
3
y
2
x
+ 2y
xx
y
xx
= 0,
2y 2y
2
x
+ y
2
y
y
2
xx
= 0,
X
1
=
x
, X
2
= x
x
2y
y
, (4.86)
comprobando que
[X
1
, X
2
] = X
1
.
De esta forma las variables cannicas que induce este campo X
1
son:
(t = y, u(t) = x) = (x = u(t), y = t) (4.87)
en estas nuevas variables la ode original (4.85) se rescribe como:
u
=
u
_
1 + t
3
u
2
_
t
(4.88)
que no parece muy fcil de resolver, por esta razn volveremos a utilizar las variables cannicas para reducir el
orden de la ode original (4.85), resultando
u
=
u
t
u
3
t
2
(4.89)
donde
_
t = y, u(t) =
1
y
_
, (4.90)
encontrando que la ecuacin (4.89) es de tipo Bernoulli y cuya solucin inmediata es:
u =
1
t
2t + C
1
, (4.91)
pero imaginemos que no sabemos resolverla. En dicho caso aplicamos otra vez el mtodo de Lie obteniendo por
lo tanto (simplemente aplicamos el procedimiento standard):
t
+ (
u
t
)
_
u
t
u
3
t
2
_
u
_
u
t
u
3
t
2
_
2
_
u
t
2
2u
3
t
_
1
t
3u
2
t
2
_
= 0
encontrando como solucin los siguientes campos:
X
1
= u
3
t
2
u
, X
2
=
_
u 2u
3
t
3
_
u
, X
3
= t
t
3
2
u
u
, X
4
=
t
u
t

u
4.3. EJEMPLOS 97
si por ejemplo elegimos X
4
para calcular las v.c. encontramos:
(r = ut, z(r) = t) =
_
t = z(r), u =
r
z(r)
_
(4.92)
de tal forma que la ecuacin (4.89) se rescribe en estas nuevas variables como:
z
=
1
r
3
= z(r) =
1
2r
2
+ C
1
deshaciendo el c.v. (y simplicando)
u(t) =
1
t
2t + C
1
,
que es la solucin (4.91) que ya habamos obtenido.
Otra forma de atacar la ecuacin (4.89) sera la de buscar un factor integrante que haga que ( (4.89)) sea
exacta, dicho factor integrante es:
=
1
u
3
t
2
.
Una vez obtenida la solucin a la ecuacin (4.89) entonces slo nos queda deshacer el c.v. (4.90) obteniendo la
siguiente ecuacin diferencial:
y
= y
_
2y + C
1
(4.93)
que es difcil de integrar (si lo fuese entonces volveramos a aplicar el mtodo de Lie) encontrando que
y =
1 + tan
_
x+C
2
2C
1
_
2
2C
2
1
(4.94)
es la solucin ms general a la ecuacin planteada (4.85).
Veremos ahora como obtener soluciones particulares. De las dos simetras obtenidas al sistema de edps i.e. (4.86)
tomamos X
2
= x
x
2y
y
, i.e. una simetra de escala y calculamos la solucin invariante que induce i.e.
dx
x
=
dy
2y
= y =
2
x
2
es una solucin particular a la ecuacin (4.85). A igual resultado podemos llegar si aplicamos nuestro mtodo
pedestre (el AD) simplemente observando que dicha ecuacin ser dimensionalmente homognea si introduci-
mos una constante dimensional i.e. la ecuacin (4.85) se rescribe como
y
=
y
2
y
+ Ky
2
(4.95)
tal que [K] = y
1
x
2
por lo tanto la solucin que nos sugiere el AD es del tipo
y K x
y 1 1 0
x 0 2 1
= y
1
Kx
2
tal y como ya sabamos por la solucin invariante.
Observacin 4.3.1 Si insistemos en utilizar la simetra de escala para obetener un cv, veremos que la cosa se complica
muchsimo. Vemos que X
2
= x
x
2y
y
induce el siguiente cv
t = yx
2
, u(t) =
1
2 (2y + y
x)
,
que nos lleva a obtener una ODE tipo Abel
u
=
_
t
2
2t
_
u
3
u
2
u
t
,
la cual s tiene solucin, el problema est en deshacer el cv en dicha solucin. Por lo tanto el cv que inducira el AD (ie. el
de la simetra de escala X
2
) NO FUNCIONA.
y
=
y
y
2

1
xy
(4.96)
y
3
yy
= 0,
2
y
+ y
2
yy
2y
2
xy
= 0,
3
y
2
x

y
2 y
x
+ 2y
3
xy
y
3
xx
= 0,
y
x
+ 2
y
2
x

x
y
2
x

y
y
2
x
y
x
+ y
3
xx
= 0,
X
1
= x
2
x
+ xy
y
, X
2
= x
x
+
y
2
y
, (4.97)
comprobando que
[X
1
, X
2
] = X
1
.
1
son:
_
t =
y
x
, u(t) =
1
x
_
=
_
x =
1
u(t)
, y =
t
u(t)
_
(4.98)
u
=
u
2
t
2
(4.99)
que parece fcil de resolver, por ejemplo podemos encontrar un factor integrante que haga que dicha ode sea
exacta, este factor integrante es:
=
1
u
2
de tal forma que la primera integral de la ode resultante es:
1
u

1
t
+ C
1
= 0
distinguiendo dos casos. Si C
1
= 0 entonces
1
u

1
t
= 0 = u(t) =
t
2
2
+ K
4.3. EJEMPLOS 99
que deshaciendo el c.v. nos lleva a obtener:
y =
_
2Kx
2
+ 2x (4.100)
mintras que si C
1
= 0 entonces
u(t) =
t
C
1
+
ln (C
1
t 1)
C
1
+ C
2
deshaciendo el c.v. obtenemos por lo tanto:
1
x
=
y
xC
1
+
ln
_
C
1
y
x
1
_
C
1
+ C
2
. (4.101)
En este milagroso ejemplo todo ha resultado extremadamente sencillo pero para demostrar como funciona el
mtodo haremos como en el ejemplo anterior y volveremos a utilizar las variables cannicas para reducir el
u
=
u
2
t
2
(4.102)
donde
_
t =
y
x
, u(t) =
1
y
x y
_
, (4.103)
encontrando que la ecuacin (4.102) es de tipo separable y cuya solucin inmediata es:
u =
t
C
1
t 1
, (4.104)
t
+ (
u
t
)
_
u
2
t
2
_
u
_
u
2
t
2
_
2
2
_
u
2
t
3
_
+ 2
_
u
t
2
_
= 0
X
1
= u
2
u
, X
2
= t
2
t
, X
3
=
u(t + u)
t

u
, X
4
=
_
1 +
u(2t + u)
t
2
_
u
X
5
= t
t
+ u
u
, etc...
4
_
r = t, z(r) =
t
2
t + u
_
=
_
t = r, u =
r (z(r) + r)
z(r)
_
(4.105)
z
= 1 = z(r) = r + C
1
u(t) =
t
C
1
t 1
,
=
1
u
2
.
y
=
y
x
+
C
1
x

1
y
(4.106)
que es difcil de integrar (es de tipo Abel) por lo tanto volvemos a aplicar el mtodo de Lie, encontrando que
x
+
_
x
_
_
y
x
+
C
1
x

1
y
_
y
_
y
x
+
C
1
x

1
y
_
2
+
_
y
x
2
+
C
1
x
2
_
_
1
x
+
1
y
2
_
= 0
encontrando slo dos simetras
X
1
=
_
x
2
y
xC
1
_
y
, X
2
= x
2
x
+ xy
y
,
el c.v. que induce X
2
= x
2
x
+ xy
y
es:
_
r =
y
x
, z(r) =
1
x
_
=
_
x =
1
z(r)
, y =
r
z(r)
_
la ecuacin (4.106) se rescribir como:
z
=
r
rC
1
1
(4.107)
que es una cuadratura
z =
r
C
1
+
ln(rC
1
1)
C
2
1
+ C
2
por ltimo deshacemos el c.v. obteniendo por lo tanto la solucin (4.101)
1
x
=
y
xC
1
+
ln
_
C
1
y
x
1
_
C
1
+ C
2
(4.108)
Veremos ahora como obtener soluciones particulares. De las dos simetras obtenidas al sistema de edps i.e. (4.97)
tomamos X
2
= x
x
+
y
2
y
dx
x
=
2dy
y
= y =
2x
es una solucin particular a la ecuacin (4.96). A igual resultado podemos llegar si aplicamos nuestro mtodo
pedestre (el AD) simplemente observando que dicha ecuacin ser dimensionalmente homognea si introduci-
mos una constante dimensional i.e. la ecuacin (4.96) se rescribe como
y
= K
1
y
y
2

K
2
xy
(4.109)
tal que [K
1
] = y
2
x
1
, [K
2
] = y
2
x
1
, por lo tanto la solucin que nos sugiere el AD es del tipo
y K
1
x
y 1 2 0
x 0 1 1
= y
Kx
y
=
y
2
y
+
_
y
1
y
_
y
(4.110)
4.3. EJEMPLOS 101
y
yy
+
y
= 0,
_
1 y
2
_
y
y
+ y
2
yy
2y
2
xy
y
y
+ = 0,
2y
xy
y
xx
2y
2
y
+
_
1 y
2
_
x
2
x
_
y
2
1
_
= 0,
_
1 y
2
_
x
+ y
xx
= 0,
X
1
=
x
, X
2
=
y
1
_
y
2
+ 1 y
1
_
y

y
, (4.111)
comprobando que
[X
1
, X
2
] = 0.
Las variables cannicas que induce este campo X
1
son:
(t = y, u(t) = x) = (x = u(t), y = t) (4.112)
en estas nuvas variables la ode original (4.110) se rescribe como:
u
=
_
1 + u
t
2
u
_
u
t
(4.113)
que no parece fcil de resolver, entonces volvemos a utilizar las variables cannicas para reducir el orden de la
ode original (4.110), resultando
u
=
_
t
2
1
_
u
2
t

u
t
(4.114)
donde
_
t = y, u(t) =
1
y
_
, (4.115)
u =
t
t
2
+ C
1
t + 1
, (4.116)
y
= y
2
+ yC
1
+ 1 (4.117)
que es fcil de integrar (es de tipo cuadratura) por lo tanto la solucin de (4.110) es
x +
2
_
_
C
2
1
4
_
arctanh
2y + C
1
_
_
C
2
1
4
_
+ C
2
= 0 (4.118)
o alternativamente:
y = C
1
_
C
2
1
1 tanh
_
_
C
2
1
1(x + C
2
)
_
si C
2
1
> 1,
y = C
1
+
_
1 C
2
1
tan
_
_
1 C
2
1
(x + C
2
)
_
si C
2
1
< 1,
y = C
1
(x + C
2
)
1
si C
2
1
= 0,
y
=
_
3
x
2x
_
y
+ 4y (4.119)
yy
= 0,
2
_
3
x
2x
_
y
+
yy
2
xy
= 0,
12y
y
xx
+
_
3
x
2
+ 2x
_

_
3
x
2x
_
x
+ 2
xy
= 0,
8y
x
4 + 4y
y
_
3
x
2x
_
x
+
xx
= 0,
X
1
= y
y
, X
2
=
_
x
2
1
_
y
, X
3
= e
x
2
y
, X
4
=
y
x
3
2y
2
x
2

y
,
X
5
=
x
2
1
x
3

x
+
2y
x
2
y
, X
6
=
e
x
2
x
3

x
2ye
x
2
x
2

y
, (4.120)
X
7
=
y(x
2
1)e
x
2
x
3

x
2y
2
e
x
2
x
2

y
, X
8
=
(x 1)
2
(x + 1)
2
e
x
2
x
3

x
2y (x 1) (x + 1) e
x
2
x
2

y
,
en este caso el lgebra que generan estos 8 campos (X
i
)
8
i=1
no es soluble
2
son:
_
t = x, u(t) =
y
(x
2
1)
_
=
_
x = t, y = u(t) + u(t)
2
t
2
_
(4.121)
u
=
_
2t
4
t
2
+ 3
_
u
t (t
2
1)
(4.122)
que parece fcil de resolver,
u
=
_
2t
4
t
2
+ 3
_
t (t
2
1)
= u
=
t
3
e
t
2
+C
1
(1 + t)
2
(1 + t)
2
=
u =
1
2
e
t
2
+C
1
t
2
1
=
C
2
e
t
2
t
2
1
+ C
1
si deshacemos el c.v. obtenemos:
y
(x
2
1)
=
C
2
e
x
2
x
2
1
+ C
1
= y = C
1
_
x
2
1
_
+ C
2
e
x
2
(4.123)
Si volvemos a utilizar las variables cannicas para reducir el orden de la ode original (4.122), resultando
u
=
_
2t
4
t
2
+ 3
_
u
t (t
2
1)
(4.124)
donde
_
t = x, u(t) =
2yx + (1 x
2
)y
(x 1)
2
(x + 1)
2
_
, (4.125)
4.3. EJEMPLOS 103
u =
t
3
e
t
2
+C
1
(1 + t)
2
(1 + t)
2
, (4.126)
y
=
x
3
e
x
2
+C
1
(x
2
1)
+
2xy
(x
2
1)
(4.127)
que es fcil de integrar (es de tipo lineal) por lo tanto la solucin de (4.119) es
y = C
1
_
x
2
1
_
+ C
2
e
x
2
(4.128)
tal y como ya sabamos.
y
= p (x) y
+ q(x) (4.129)
yy
= 0,
2p(x)
y
+
yy
2
xy
= 0,
3yq(x)
y
xx
+ p
(x) p (x)
x
+ 2
xy
= 0,
2q(x)
x
+ q(x)
y
p(x)
x
+
xx
q
(x) = 0,
X
1
=
y
, X
2
=
_
e
_
p(x)dx
dx
y
, (4.130)
1
son:
(t = x, u(t) = y) = (x = t, y = u(t)) (4.131)
u
= p (t) u
+ q(t) (4.132)
que parece fcil de resolver,
u =
_
e
_
p(t)dt
_
_
q(t)e
_
p(t)dt
dt + C
1
_
dt + C
2
si deshacemos el c.v. obtenemos:
y =
_
e
_
p(x)dx
_
_
q(x)e
_
p(x)dx
dx + C
1
_
dx + C
2
(4.133)
Si volvemos a utilizar las variables cannicas para reducir el orden de la ode original (4.132), resultando
u
= p(t)u + q(t) (4.134)

donde
_
t = x, u(t) = y
_
, (4.135)
encontrando que la ecuacin (4.134) es de tipo lineal y cuya solucin inmediata es:
u =
_
_
q(t)e
_
p(t)dt
dt + C
1
_
e
_
p(t)dt
, (4.136)
y
=
_
_
q(x)e
_
p(x)dx
dx + C
1
_
e
_
p(x)dx
, (4.137)
que es fcil de integrar (cuadraturas) por lo tanto la solucin de (4.129) es
y =
_
e
_
p(x)dx
_
_
q(x)e
_
p(x)dx
dx + C
1
_
dx + C
2
(4.138)
tal y como ya sabamos.
y
=
y
x
+
3y
2
2x
3
, (4.139)
tipo Emden-Fowler.
x
2
yy
= 0,
2x
y
+ x
2
yy
2x
2
xy
= 0,
9y
2
2x

y
x
x
+ + 2x
2
xy
x
2
xx
= 0,
3y
x

3y
2
x

x
+
3y
2
2x

y
+
9y
2
2x
3
x
x
+ x
2
xx
= 0,
cuya solucin es:
X
1
= x
x
+ y
y
. (4.140)
1
son:
_
t =
y
x
, u(t) = ln x
_
=
_
x = e
u(t)
, y = te
u(t)
_
(4.141)
u
=
_
u
_
3
t
_
1 +
3
2
t
_
(4.142)
que parece muy fcil de resolver,
u
(u
)
3
= t
_
1 +
3
2
t
_
=
1
2u
2
=
1
2
t
3
1
2
t
2
+ C
1
(4.143)
4.3. EJEMPLOS 105
por lo tanto
u
=
dt
2
_
1
2
t
3
+
1
2
t
2
+ C
1
= u =
_
dt
_
2t
3
+ 2t
2
+ C
1
(4.144)
si C
1
= 0 entonces encontramos una solucin particular:
u =
2 arctanh
1
2
_
(2t + 2)
2 (4.145)
as que si deshacemos el c.v. obtenemos:
ln x =
_
y
x da
_
2a
3
+ 2a
2
+ C
1
+ C
2
. (4.146)
Por esta razn volveremos a utilizar las variables cannicas para reducir el orden de la ode original (4.139),
resultando
u
=
1
2
u
3
t (2 + 3t) (4.147)
donde
_
t =
y
x
, u(t) =
x
y xy
_
, (4.148)
u =
1
_
t
3
+ t
2
+ C
1
, (4.149)
x
y xy
=
1
_
y
2
x
2
+
y
3
x
3
+ C
1
(4.150)
que no es difcil de integrar, encontrando que
ln x =
_
y
x da
_
2a
3
+ 2a
2
+ C
1
+ C
2
. (4.151)
Veremos ahora como obtener soluciones particulares. De las simetras obtenidas al sistema de edps i.e. (4.140)
tomamos X
1
= x
x
+ y
y
dx
x
=
dy
y
= y = ax
es una solucin particular a la ecuacin (4.139) siendo a R en particular encontramos que a = 2/3. A igual
resultado podemos llegar si aplicamos nuestro mtodo pedestre (el AD) simplemente observando que dicha
ecuacin ser dimensionalmente homognea si introducimos una constante dimensional i.e. la ecuacin (4.139)
se rescribe como
y
=
y
2
x
+ K
3y
2
2x
3
(4.152)
tal que [K] = y
1
x por lo tanto la solucin que nos sugiere el AD es del tipo
y K x
y 1 1 0
x 0 1 1
= y
x
K
tal y como ya sabamos por la solucin invariante. Esta transformacin nos llevara a obtener el mismo cv que
el obetenido mediante GL.
y
=
y
x
+
3
2
y
2
x
3
, (4.153)
tipo Emden-Fowler.
Solucin. Aplicando el mtodo de Lie lo primero que tenemos que plantear el es sistema de edps que nos
x
2
yy
= 0,
2x
y
+ x
2
yy
2x
2
xy
= 0,
9
2
y
2
x
5
y
x
x
+ + 2x
2
xy
x
2
xx
= 0,
3y
2
x
5
3y
2
x
5
x
+
3
2
y
2
x
5
y
+
9
2
y
2
x
4
x
x
+ x
2
xx
= 0,
cuya solucin es:
X
1
= x
x
5y
y
. (4.154)
1
son:
_
t = x
5
y, u(t) = ln x
_
=
_
x = e
u(t)
, y = te
5u(t)
_
(4.155)
u
= 12u
2
+ 35u
3
3
2
u
3
t
2
(4.156)
que no parece muy fcil de resolver.
Por esta razn volveremos a utilizar las variables cannicas para reducir el orden de la ode original (4.153),
resultando
u
= u
3
_
35t
3
2
t
2
_
12u
2
(4.157)
donde
_
t =
y
x
, u(t) =
x
y xy
_
, (4.158)
encontrando que la ecuacin (4.157) es de tipo Abel y por desgracia es de las que no sabemos resolver. Si
aplicamos el mtodo de Lie a dicha ecuacin veramos que no somos capaces de encontrar simetra alguna.
Por esto la solucin a la que podemos llegar ode original (4.153) es del tipo:
y =
a
e
_
b(a)da+C
1
, (4.159)
donde
b
= b
3
_
35a
3
2
a
2
_
12b
2
, b = b(a). (4.160)
Veremos ahora como obtener soluciones particulares. De las simetras obtenidas al sistema de edps i.e. (4.154)
tomamos X
1
= x
x
5y
y
dx
x
=
dy
5y
= y =
a
x
5
es una solucin particular a la ecuacin (4.153) siendo a R en particular encontramos que a = 70/3. A igual
resultado podemos llegar si aplicamos nuestro mtodo pedestre (el AD) simplemente observando que dicha
4.3. EJEMPLOS 107
ecuacin ser dimensionalmente homognea si introducimos una constante dimensional i.e. la ecuacin (4.153)
se rescribe como
y
=
y
2
x
+ K
3
2
y
2
x
3
(4.161)
tal que [K] = y
1
x por lo tanto la solucin que nos sugiere el AD es del tipo
y K x
y 1 1 0
x 0 5 1
= y
1
Kx
5
y
=
y
x
e
y
(4.162)
x
2
yy
= 0,
2x
y
+ x
2
yy
2x
2
xy
= 0,
3x
2
e
y
y
+ x
x
+ 2x
2
xy
x
2
xx
= 0,
x
2
e
y
+ 2x
2
e
y
x
x
2
e
y
y
+ x
x
+ x
2
xx
= 0,
X
1
= x
x
2
y
, X
2
= x ln x
x
2 (1 + ln x)
y
, (4.163)
comprobando que
[X
1
, X
2
] = X
1
.
1
son:
(t = y + ln x, u(t) = ln x) =
_
x = e
u(t)
, y = t 2u(t)
_
(4.164)
u
=
(u
)
3
e
(t2u)
e
2u
(4.165)
u
= u
3
e
t
(4.166)
donde
_
t = y + 2 ln x, u(t) =
1
2 + xy
_
, (4.167)
u =
1
_
2e
t
+ C
1
, (4.168)
t
+ (
u
t
)
_
u
3
e
t
_
u
_
u
3
e
t
_
2
_
u
3
e
t
_
3
_
u
2
e
t
_
= 0
X
1
= u
3
u
, X
2
=
_
u + 2u
3
e
t
_
u
, X
3
=
t
u
2
u
, X
4
= e
t
t
3
_
r = ue
t/2
, z(r) = t
_
=
_
t = z(r), u =
r
e
1/2z(r)
_
(4.169)
z
=
2
r (2r
2
+ 1)
= z(r) = ln
_
2r
2
+ 1
_
+ 2 lnr + C
1
t = ln
_
2
_
ue
t/2
_
2
+ 1
_
+ 2 ln
_
ue
t/2
_
+ C
1
,
=
1
u
3
.
y
=
2 +
_
2e
(y+2 ln x)
+ C
1
x
(4.170)
que es difcil de integrar, por eso volvemos a aplicar el mtodo de Lie encontrando que
x
+
_
x
_
_
2 +
_
2e
(y+2 ln x)
+ C
1
x
_
y
_
2 +
_
2e
(y+2 ln x)
+ C
1
x
_
2
+
+
_
2
_
(2e
y
x
2
+ C
1
) + C
1
_
(2e
y
x
2
+ C
1
)x
2
_
+ 3
_
xe
y
_
(2e
y
x
2
+ C
1
)
_
= 0
X
1
= x
x
2
y
, X
2
=
_
2x
2
e
y
+ C
1
y
,
1
(r = y + 2 ln x, z(r) = ln x) = (x = e
z
, y = r 2z) (4.171)
z
=
1
2e
r
+ C
1
= z(r) =
2arctanh
_
2e
r
+C
1
C
1
_
C
1
+ C
2
4.3. EJEMPLOS 109
ln x =
2arctanh
_
2e
y
x
2
+C
1
C
1
_
C
1
+ C
2
(4.172)
que es la solucin de (4.170) y por lo tanto de (4.162).
y
yx
6
2y
3
x
6
+ 2y
2
yx
5
+ y
5
= 0 (4.173)
y
yy
+ 2
y
= 0,
y
2
x
yy
+ 2x 2y
2
x
xy
+ 4y
2
y
2yx
y
= 0,
2y
2
x
6
x
4yx
7
x
+ 2y
2
x
7
xy
y
2
x
7
xx
2y
2
x
5
= 0,
2y
2
x
6
x
+ y
2
x
7
xx
+ 4y
6
x
y
= 0,
6y
6
+ 4xy
5
+ 3y
6
x
x
2y
6
x
y
= 0,
y
6
x
x
= 0,
X
1
= y
2
y
, X
2
= x
2
x
, X
3
= x
x
+
3
2
y
y
, (4.174)
comprobando que
X
1
X
2
X
3
X
1
0 0
3
2
X
1
X
2
0 0 X
2
X
3
3
2
X
1
X
2
0
i.e.
_
X
i
, X
j
= X
k
1
son:
_
t = x, u(t) =
1
y
_
=
_
x = t, y =
1
u
_
(4.175)
u
=
2u
t

1
u
t
6
(4.176)
u
=
2u
t

1
ut
6
(4.177)
donde
_
t = x, u(t) =
y
y
2
_
, (4.178)
u =
_
2t + t
2
C
1
t
3
, (4.179)
y
=
y
2
_
2x + x
2
C
1
x
3
(4.180)
que no es difcil de integrar, encontrando que
1
y

_
2x + x
2
C
1
_
3/2
3x
3
C
2
= 0 (4.181)
y
+ xy
+ xy
3
y yy
2
= 0 (4.182)
y
y
x
x
+ 2x
y
yy
= 0,
y
y
+ 2x
y
+ +3x
x
+
yy
2y
2
xy
= 0,
3y
y
+ x
x
2y
x
+ 2
xy
xx
+ = 0,
2y
x
+
xx
+ y
y
+ x
x
= 0,
cuya solucin es:
X
1
= y
x
x
y
, (4.183)
1
son:
_
t = x
2
+ y
2
, u(t) = arctan
_
x
y
__
(4.184)
u
=
1
2
u
_
4t
2
u
2
+ 4t
3
u
2
+ t + 3
_
t
(4.185)
u
= 2u
3
t (1 + t)
u (t + 3)
2t
(4.186)
donde
_
t = x
2
+ y
2
, u(t) =
xy
y
2 (x
2
+ y
2
) (x + yy
)
_
, (4.187)
u =
1
t
_
4 + te
t
+ C
1
, (4.188)
4.3. EJEMPLOS 111
xy
y
(x + yy
)
=
2
_
4 + (x
2
+ y
2
) e
(x
2
+y
2
)
+ C
1
(4.189)
i.e.
y
=
2x + y
_
4 + (x
2
+ y
2
) e
(x
2
+y
2
)
+ C
1
_
x
_
4 + (x
2
+ y
2
) e
(x
2
+y
2
)
+ C
1
2y
_ (4.190)
que es difcil de integrar, encontrando que
arctan
_
x
y
_
_
x
2
+y
2
1
a
4 + ae
a
+ C
1
da C
2
= 0 (4.191)
Si aplicamos el mtodo de Lie a la ecuacin (4.190) encontraramos que una de las simetras que admite dicha
ecuacin es X
2
= y
x
+ x
y
que a suvez induce un c.v.
_
r = x
2
+ y
2
, z(r) = arctan
_
x
y
__
(4.192)
de tal forma que (4.190) se escribe como
z
=
1
r
4 + e
r
+ rC
1
(4.193)
que es una ode tipo cuadratura
z =
_

1
r
4 + e
r
+ rC
1
dr + C
2
(4.194)
por lo tanto deshaciendo el c.v. llegaramos a obtener la solucin (4.191).
y
=
y
2
x
2

2yy
x
+ y
2
(4.195)
Solucin. Las simetras que admite la ODE son
X
1
= x
x
, X
2
= x
y
, (4.196)
por lo que
[X
1
, X
2
] = X
2
.
2
son:
t = x, u(t) =
y
x
, (4.197)
en estas nuevas variables la ode original se rescribe como:
u
=
u
_
u
t
2
2
_
t
(4.198)
y cuya solucin es:
u =
_

dt
t
2
(ln t C
1
)
+ C
2,
i.e.
y
x
=
_

dx
x
2
(ln x C
1
)
+ C
2,
Para reducir la ODE utilizamos las variables cannicas, resultando
u
= u
2
t
2u
t
, (4.199)
de tipo Bernoulli donde
t = x, u(t) =
y
x y
x
2
, (4.200)
y cuya solucin inmediata es:
u =
1
t
2
(ln t C
1
)
, (4.201)
Una vez obtenida la solucin a la ecuacin de Bernoulli, u(t), entonces slo nos queda deshacer el c.v. obteniendo
la siguiente ecuacin diferencial:
y
x y
x
2
=
1
x
2
(ln x C
1
)
(4.202)
i.e.
y
=
1
x (ln x C
1
)
+
y
x
(4.203)
que no es difcil de integrar (lineal), encontrando que
y =
_
_

dx
x
2
(ln x C
1
)
+ C
2
_
x (4.204)
es la solucin ms general a la ecuacin planteada.
Chapter 5
Ecuaciones de tercer orden.
5.1 Ejemplos
Veamos algunos ejemplos d eecuaciones de tercer orden y de como funciona el mtodo con ellas.
Ejemplo 5.1.1 Resolver la siguiente ecuacin de tercer orden
y
=
1
y
3
(5.1)
y
3
y
= 0,
y
3
yy
= 0,
9y
3
xy
+ 3y
3
yy
= 0,
3y
3
xy
3y
3
xx
= 0,
y
3
yyy
= 0,
3y
3
xyy
+ y
3
yyy
= 0,
3y
3
xxy
+ y
3
xyy
= 0,
4
y
+ 3y
3
xxy
3y
3
xxx
= 0,
3
y
+
y
3
x
+ y
3
xxx
= 0,
X
1
=
x
, X
2
= x
x
+
3
4
y
y
, (5.2)
comprobando que
[X
1
, X
2
] = X
1
vemos que al encontrar slo 2 simetras, entonces nuestra ode no va a poder ser reducida a una ode tipo cuadrat-
uras.
113
114 CHAPTER 5. ECUACIONES DE TERCER ORDEN.
1
son:
(t = y, u(t) = x) = (x = u, y = t) (5.3)
u
=
3t
3
u
2
u
5
t
3
u
(5.4)
u
=
2u
2
u

u
4
t
3
(5.5)
donde
_
t = y, u(t) =
1
y
_
, (5.6)
encontrando que la ecuacin (5.5) es algo compicadilla de resolver as que aplicamos d enuevo el mtodo de Lie
i.e.
3u
u
u
2
uu
= 0,
3 3u
u
+ u
2
uu
2u
2
tu
= 0,
3
u
6
t
3

u
6u
t
+ 2u
2
tu
u
2
tt
= 0,
4
u
5
t
3
2
u
6
t
3

t
u
6
t
3

u
3
u
6
t
4
+ u
2
tt
= 0,
comprobando que admite una nica simetra
X
3
= 3t
t
+ u
u
(5.7)
que induce el siguiente c.v. que nos permite obtener una ode de primer orden:
s
=
1
9
_
9r
5
5r
2
_
s
3
r

7
3
s
2
3s
r
(5.8)
donde
_
r =
u
t
1/3
, s(r) =
3t
1/3
u 3u
t
_
(5.9)
que por desgracia al ser de tipo Abel no sabemos como resolverla.
Intentaremos encontrar una solucin particular mediante el mtodo de los invariantes. La solucin que induce
el campo X
2
= x
x
+
3
4
y
y
es por lo tanto:
dx
x
=
4dy
3y
= y = ax
3/4
(5.10)
donde a R, un simple clculo nos nuestra que
a =
2
2 (15)
3/4
15
El mtodo pedestre nos lleva a obtener la misma solucin despus de homogeneizar dimensionalmente la ode
mediante la introduccin de una constante dimensional
y
=
K
y
3
(5.11)
5.1. EJEMPLOS 115
donde [K] = y
4
x
3
, por lo tanto
y K x
y 1 4 0
x 0 3 1
= y
_
Kx
3
_
1/4
(5.12)
como ya sabamos.
y
= yy
(5.13)
y
= 0,
yy
= 0,
9
xy
+ 3
yy
+ y
y
= 0,
+ y
x
3
xx
+ 3
xy
= 0,
yyy
= 0,
3
xyy
+
yyy
y
yy
= 0,
3
xyy
3
xxy
+ y
yy
2y
xy
= 0,
xxx
+ 3
xxy
y
xx
+ 2y
xy
= 0,
xxx
+ y
xx
= 0,
X
1
=
x
, X
2
= x
x
y
y
, (5.14)
comprobando que
[X
1
, X
2
] = X
1
vemos que al encontrar slo 2 simetras, entonces nuestra ode no va a poder ser reducida a una ode tipo cuadrat-
uras.
1
son:
(t = y, u(t) = x) = (x = u, y = t) (5.15)
u
=
u
_
tu
2
3u
_
u
(5.16)
u
=
3u
2
u
u
ut (5.17)
donde
_
t = y, u(t) =
1
y
_
, (5.18)
encontrando que la ecuacin (5.17) es algo compicadilla de resolver as que aplicamos d enuevo el mtodo de
Lie i.e.
3u
u
u
2
uu
= 0,
3 + 2u
3
t
u
3u
u
+ u
2
uu
2u
2
tu
= 0,
u
2
t + u
3
t
t
+ u
3
6u
t
+ 2u
2
tu
u
2
tt
= 0,
u
3
t
t
+ u
2
tt
= 0,
comprobando que admite una nica simetra
X
3
= t
t
2u
u
(5.19)
s
=
_
2r
3
+ 6r
2
_
s
3
r
+
_
r
2
7r
_
s
2
r

3s
r
(5.20)
donde
_
r = ut
2
, s(r) =
1
t
2
(u
t + 2u)
_
(5.21)
que por desgracia al ser de tipo Abel no sabemos como resolverla.
el campo X
2
= x
x
y
y
es por lo tanto:
dx
x
=
dy
y
= y =
a
x
(5.22)
donde a R, un simple clculo nos nuestra que a = 3.
y
= Kyy
(5.23)
donde [K] = y
1
x
1
, por lo tanto
y K x
y 1 1 0
x 0 1 1
= y
1
Kx
(5.24)
como ya sabamos.
y
=
y
2
y
(1 + y
)
(5.25)
5.1. EJEMPLOS 117
y
= 0,
x
+
y
2
y
= 0,
x
= 0,
yy
= 0,
10
yy
9
xy
+ 3
yy
= 0,
4
yy
3
xx
+ 3
xy
14
xy
4
yy
= 0,
2
xy
4
xx
5
xy
yy
= 0,
2
xx
xy
xx
= 0,
xx
= 0,
yyy
= 0,
3
xyy
2
yyy
+
yyy
= 0,
6
xyy
3
xxy
+ 3
xyy
+ 2
yyy
yyy
= 0,
6
xxy
+ 6 + 3
xyy
xxx
+ 3
xxy
3
xyy
+
yyy
= 0,
6
xxy
2
xxx
+
xxx
3
xxy
+ 3
xyy
= 0,
2
xxx
xxx
+ 3
xxy
= 0,
xxx
= 0,
X
1
=
y
, X
2
=
x
, X
3
= x
x
+ y
y
, (5.26)
comprobando que
[X
1
, X
2
] = X
1
vemos que al encontrar 3 simetras, entonces nuestra ode va a poder ser reducida a una ode tipo cuadraturas.
1
son:
(t = y, u(t) = x) = (x = u, y = t) (5.27)
u
=
u
2
(2u
+ 3)
u
(u
+ 1)
(5.28)
En esta ocasin utilizaremos las variables que cannicas que inducen cada una de las simetras obtenidas hasta
llegar a una solucin (en este caso podemos), resultando
u
=
u
2
(2u + 3)
u (u + 1)
(5.29)
donde
_
t = y, u(t) =
1
y
_
, (5.30)
la siguiente reduccin es:
s
=
s (2r + 3)
r (r + 1)
(5.31)
donde
_
r = u(t), s(r) =
1
u
_
(5.32)
por lo tanto
s = e
_
h(i)di
, h(i) =
(2i + 3)
i (i + 1)
(5.33)
donde
_
i = r, h(i) =
s
s
_
(5.34)
de esta forma, deshaciendo todos los cambios de variables y resolviendo las odes resultantes llegamos a la
solucin de la ode original (5.25) resultando ser:
x =
_
1 2C
1
y 2C
1
C
2
C
1
1
2
ln
__
1 2C
1
y 2C
1
C
2
1
_
C
1
+
+
1
2
ln
__
1 2C
1
y 2C
1
C
2
+ 1
_
C
1
1
2
ln (y + C
2
)
C
1
+ C
3
(5.35)
este es en realidad el camino que deberamos seguir en cada caso, pero he preferido ser especialmente pesado
desarrollando otras tcticas ya adquiridas.
El camino que siempre hemos seguido sera el de estudiar la ode (5.29) aplicando de nuevo el mtodo de Lie i.e.
planteando un nuevo sistema de edps para buscar las simetras de dicha ecuacin, en este caso resulta que dicha
ecuacin admite muchas simetras entre ellas X
4
= (u
2
+ u)
u
X
4
= (u
2
+ u)
u
(5.36)
s
= 2s
2
(5.37)
donde
_
r = t, s(r) =
u
u (u + 1)
_
(5.38)
por lo tanto
s =
1
2r C
1
(5.39)
y deshaciendo el c.v.
u
u (u + 1)
=
1
2t C
1
=
1
u
= C
2
_
2t C
1
1 (5.40)
de nuevo y por ltimo deshacemos el ltimo c.v. obtenindose:
y
= C
2
_
2y C
1
1 (5.41)
y cuya solucin es precisamente la ya encontrada i.e. la solucin (5.35).
el campo X
3
= x
x
+ y
y
es por lo tanto:
dx
x
=
dy
y
= y = ax (5.42)
donde a R.
y
+ Ky
2
y
= y
2
(5.43)
donde [K] = y
1
x, por lo tanto
y K x
y 1 1 0
x 0 1 1
= y
x
K
(5.44)
5.1. EJEMPLOS 119
como ya sabamos.
y
=
2y
2
y
+
y
x
+
y
2
x
(5.45)
x
2
y
= 0,
x
2
x
= 0,
x
2
yy
= 0,
x
2
yx
x
2
yy
x
y
= 0,
5x
2
xy
x
x
+ x
2
xx
= 0,
x
2
xx
= 0,
x
2
yyy
= 0,
2x
y
+ x
2
yyy
3x
2
xyy
+ x
yy
= 0,
x
y
x
x
+ 3x
2
xyy
3x
2
xxy
+ 2x
xy
x
yy
= 0,
2x
x
x
2
xxx
+ 3x
2
xxy
2x
xy
+ x
xx
= 0,
x
xx
+ x
2
xxx
= 0,
X
1
=
y
, X
2
= x
x
. (5.46)
1
son:
(t = x, u(t) = y) = (x = t, y = u) (5.47)
u
=
2u
2
t + u
+ u
3
u
t
(5.48)
En esta ocasin utilizaremos las variables que cannicas que inducen cada una de las simetras obtenidas hasta
llegar a una solucin (en este caso podemos), resultando
u
=
2u
2
u
+
u
t
+
u
2
t
(5.49)
donde
_
t = x, u(t) = y
_
, (5.50)
aplicando de nuevo el mtodo de Lie i.e. planteando un nuevo sistema de edps para buscar las simetras de
dicha ecuacin, en este caso resulta que dicha ecuacin admite muchas simetras entre ellas
X
3
= t
t
+ u
u
(5.51)
que induce el siguiente c.v. que nos permite obtener una ode de primer orden (tipo Bernoulli):
s
=
_
r
3
r
2
_
s
3
r

2s
r
(5.52)
donde
_
r = ut, s(r) =
1
t (u + u
t)
_
(5.53)
por lo tanto
s =
1
r
_
2r + 1 + r
2
C
1
(5.54)
y deshaciendo el c.v.
1
t (u + u
t)
=
1
ut
_
2ut + 1 + u
2
t
2
C
1
(5.55)
que resulta ser una ode de tipo homognea, cuya solucin es:
ln t C
2
arctan h
_
ut 1
_
2ut + 1 + u
2
t
2
C
1
_
= 0 (5.56)
de nuevo y por ltimo deshacemos el ltimo c.v. (5.50) obtenindose:
ln x C
2
arctan h
_
y
x 1
_
2y
x + 1 + (y
x)
2
C
1
_
= 0 (5.57)
y cuya solucin es:
y =
2 arctan h
_
2xC
2
1
4C
1
C
2
2
+1
_
_
4C
1
C
2
2
+ 1
+ C
3
(5.58)
Chapter 6
Aplicaciones en problemas cosmolgicos.
6.1 Distintas Aplicaciones
6.1.1 Lie method for the H-equation.
H + K
1
H
H + K
2
H
1
H
2
+ K
3
H
3
= 0
(6.1)
we consider
H =
_
K
1
H
H + K
2
H
1
H
2
+ K
3
H
3
_
(6.2)
i.e. we have an eq. type
H = f (t, H,

H) where
f (t, H,

H) =
_
K
1
H
H + K
2
H
1
H
2
+ K
3
H
3
_
(6.3)
therefore
f
t
= 0 (6.4)
f
H
=
_
K
1
H K
2
H
2
H
2
+ 3K
3
H
2
_
(6.5)
f

H
=
_
K
1
H + 2K
2
H
1

H
_
(6.6)
According with the previous theory, a vector eld X
X = (t, H)
t
+ (t, H)
H
(6.7)
is a symmetry of (6.2) iff
f
t
f
H
+
tt
+ (2
tH

tt
) H
+ (
HH
2
tH
) H
2
HH
H
3
+ ...
... +
_
H
2
t
3H
H
_
f
_
t
+ (
H

t
) H
H
2
H
_
f
H
= 0 (6.8)
121
122 CHAPTER 6. APLICACIONES EN PROBLEMAS COSMOLGICOS.
simpliying it is obtained:
f
t
= 0
f
H
=
_
K
1
H K
2
H
2
H
2
+ 3K
3
H
2
_
= K
1
H K
2
H
2
H
2
+ 3K
3
H
2
tt
, (2
tH

tt
) H
, (
HH
2
tH
) H
2
,
HH
H
3
H
f =
H
_
K
1
H
H + K
2
H
1
H
2
+ K
3
H
3
_
=
H
K
1
H
H
H
K
2
H
1
H
2
H
K
3
H
3
2
t
f = 2
t
_
K
1
H
H + K
2
H
1
H
2
+ K
3
H
3
_
= 2
t
K
1
H
H + 2
t
K
2
H
1
H
2
+ 2
t
K
3
H
3
3
H
H
f = 3H
H
_
K
1
H
H + K
2
H
1
H
2
+ K
3
H
3
_
= 3
H
K
1
H
H
2
+ 3
H
K
2
H
1
H
3
+ 3
H
K
3
H
3

H
t
+ (
H

t
) H
H
2
H
_
f
H
t
f
H
=
t
_
K
1
H + 2K
2
H
1

H
_
=
t
K
1
H +
t
2K
2
H
1

H
Hf
H
=
H
_
K
1
H

H + 2K
2
H
1

H
2
_
=
H
K
1
H

H +
H
2K
2
H
1

H
2
t

Hf
H
=
t
_
K
1
H

H + 2K
2
H
1

H
2
_
=
t
K
1
H

H
t
2K
2
H
1

H
2
H
2
H
f
H
=
H
_
K
1
H

H
2
+ 2K
2
H
1

H
3
_
=
H
K
1
H

H
2
H
2K
2
H
1

H
3
therefore
f
H
+
tt
+ (2
tH

tt
) H
+ (
HH
2
tH
) H
2
HH
H
3
+ ...
... +
_
H
2
t
3H
H
_
f
_
t
+ (
H

t
) H
H
2
H
_
f
H
= 0 (6.9)
6.1. DISTINTAS APLICACIONES 123
K
1

H K
2
H
2

H
2
+ 3K
3
H
2
+
tt
+ 2
tH

H
tt

H+
HH

H
2
2
tH

H
2
HH

H
3
H
K
1
H

H
H
K
2
H
1

H
2
H
K
3
H
3
+
2
t
K
1
H

H + 2
t
K
2
H
1

H
2
+ 2
t
K
3
H
3
+ 3
H
K
1
H

H
2
+ 3
H
K
2
H
1

H
3
+ 3
H
K
3
H
3

H+
t
K
1
H +
t
2K
2
H
1

H +
H
K
1
H

H +
H
2K
2
H
1

H
2
t
K
1
H

H
t
2K
2
H
1

H
2
H
K
1
H

H
2
H
2K
2
H
1

H
3
separating with respect to powers of

H we obtain:
1. For

H
3
HH

H
3
+ 3
H
K
2
H
1

H
3
H
2K
2
H
1

H
3
= 0
HH
+
H
K
2
H
1
= 0 (6.10)
2. For

H
2
K
2
H
2
+
HH
2
tH

H
K
2
H
1
+ 2
t
K
2
H
1
+ 3
H
K
1
H +
H
2K
2
H
1
t
2K
2
H
1
H
K
1
H
K
2
H
2
+
HH
2
tH
+
H
K
2
H
1
+ 2
H
K
1
H = 0 (6.11)
3. For

H
K
1

H + 2
tH

H
tt

H
H
K
1
H

H + 2
t
K
1
H

H + 3
H
K
3
H
3

H
+
t
2K
2
H
1

H +
H
K
1
H

H
t
K
1
H

H
K
1
+ 2
tH

tt
+
t
K
1
H + 3
H
K
3
H
3
+ 2
t
K
2
H
1
= 0 (6.12)
4. For

H
0
3K
3
H
2
+
tt
H
K
3
H
3
+ 2
t
K
3
H
3
+
t
K
1
H = 0 (6.13)
Therefore we obtain the overdetermined system:
HH
+
H
K
2
H
1
= 0 (6.14)
K
2
H
2
+
HH
2
tH
+
H
K
2
H
1
+ 2
H
K
1
H = 0 (6.15)
K
1
+ 2
tH

tt
+
t
K
1
H + 3
H
K
3
H
3
+ 2
t
K
2
H
1
= 0 (6.16)
3K
3
H
2
+
tt
H
K
3
H
3
+ 2
t
K
3
H
3
+
t
K
1
H = 0 (6.17)
Solving (6.29-6.17), we nd that
(t, H) = at + b, (t, H) = aH (6.18)
being a and b numerical constant. For this form of and eq. (6.2) admits a single symmetry
X = (at b)
t
(aH)
H
, (6.19)
the knowledge of one symmetry X might suggest the form of a particular solution as an invariant of the operator
X, i.e. the solution of
dt
(t, )
=
d
(t, )
(6.20)
This particular solution is known as an invariant solution (generalization of similarity solution). In this case
H =
1
at b
, (6.21)
which imply that
f = (at b)
1
a
that is to said we again obtain a power law solution.
6.1.2 Fluidos perfectos con constantes variables
We will use the eld equations in the form:
R
ij
1
2
g
ij
R =
8G(t)
c
4
(t)
T
ij
+(t)g
ij
(6.22)
Where the energy momentum tensor is:
T
ij
= ( + p) u
i
u
j
pg
ij
(6.23)
where p = in such a way that [0, 1] .
The cosmological equations are now:
2
f

f
+
( f

)
2
f
2
=
8G(t)
c(t)
2
p + c(t)
2
(t) (6.24)
3
( f

)
2
f
2
=
8G(t)
c(t)
2
+ c(t)
2
(t) (6.25)
where f stands for the scale factor. Applying the covariance divergence to the second member of equation (6.22)
we get:
T
j
i;j
=
_
4c
,j
c

G
,j
G
_
T
j
i

c
4
(t)
j
i
,j
8G
(6.26)
that simplies to:
+ 3( + 1)H
. .
T
j
i;j
=

c
4
8G

G
G
+ 4
c
c
. .
(6.27)
where H stands for the Hubble parameter (H =
f
f
). The last equation may be written in the form:
+ 3( + 1)H +

c
4
8G
+
G
G
4
c
c
= 0 (6.28)
or the equivalent
+ 3( + 1)H +

c
4
8G
+
G
G
4
c
c
= 0 (6.29)
In this equation we can take a general energy density or use the black body equation of state = a
4
where
a =

2
k
4
B
15c
3
h
3
so that equation (6.29) is now:
4
3
_
c
c
+
h
h
_
+ 3( + 1)H +
15
c
7
h
3
8
3
Gk
4
B
4
+
G
G
4
c
c
= 0 (6.30)
Simplifying hypothesis
Hypothesis number one (H1). We introduce
G
c
2
= const. B (6.31)
with [B] = LM
1
. In this way we obtain directly a term that controls the behavior of the energy density.
=
b
Bt
2
b R (6.32)
If we take =
b
Bt
2
and introduce = a
4
we arrive at
b
Bt
2
= a
4
= =
_
b
Ba
_
1/4
t
1/2
(6.33)
Hypothesis number two (H2) We can make a previous hypothesis on the behavior of the Plancks constant as
follows:
c h = const. or equivalently h =
A
c
, where A is a proportionality constant. We are motivated to this relation
because it leaves the radiation constant a as constant with a constant k
B
= const. (see [?])
h = Fc where [F] = L
1
M
1
. This hypothesis is introduced by some author to explain a possible variation of
the ne structure constant. We will see that such hypothesis is irreconciliable with our model (see [?]).
Hypothesis number three (H3) As a mathematical possibility (self-similar relation), obtained by means of the
similarity group acting on the equations, we will assume the following behavior for the cosmological constant:
=
d
c
2
(t)t
2
(6.34)
where d R. This hypothesis is a strong one since in some way it forces to the relation f (t) = c(t)t, but this is a
desired relation to get the causality principle preserved.
With this hypothesis we now deal with the equations
Lie methods for a perfect uid model.
The eld equations are:
2
f
f
+
_
f
f
_
2
=
8G
c
2
p +c
2
, (6.35)
3
_
f
f
_
2
=
8G
c
2
+c
2
, (6.36)
+ 3 ( + 1) H = 0, (6.37)
c
4
8G

G
G
+ 4
c
c
= 0, (6.38)
from (6.85) (6.86) it is obtained
2
f
f
2
_
f
f
_
2
=
8G
c
2
(p + ) , (6.39)
it is observed that
2
f
f
2
_
f
f
_
2
= 2 (H)
(6.40)
then
2 (H)
=
8G
c
2
(p ++ ) (6.41)
from (6.87) we can obtain
H =

3 (( + 1) )
(6.42)
therefore
_
= 12 ( + 1)
2
G
c
2
(6.43)
making 12 ( + 1)
2
= A then
_
= A
G
c
2
(6.44)
expanding we obtain
=

2
+ A
G
c
2
2
. (6.45)
A vector eld X
X = (t, )
t
+ (t, )
(6.46)
f
t
f
+
tt
+
_
2
t
tt
_
+
_
2
t
_
3
+ ...
... +
_
2
t
3
_
f
_
t
+
_
t
_
y
2
_
f
y
= 0 (6.47)
By expanding and separating (6.92) with respect to powers of
we obtain the overdetermined system:
+
1
= 0 (6.48)
2
t
+
2

1
= 0 (6.49)
2
t
tt
3A
G
c
2
2
1
t
= 0 (6.50)
tt
A
_
G
c
2
2G
c
c
3
_
2
2A
G
c
2
+
_
2
t
_
A
G
c
2
2
= 0 (6.51)
(t, ) = 2et + a, (t, ) = (bt + d) (6.52)
subject to the constrain
G
G
= 2
c
c
+
bt + d 4e
2et a
(6.53)
with a, b, e, and d constants. In order to solve (6.113) we can consider the following cases:
1. Case I: b = 0 and d 4e = 0 that brings us to the solution
G
G
= 2
c
c
=
G
c
2
= B = const. (6.54)
The knowledge of one symmetry X might suggest the form of a particular solution as an invariant of the
operator X i.e. the solution of
dt
(t, )
=
d
(t, )
(6.55)
this particular solution is known as an invariant solution (generalization of similarity solution), therefore
the energy density is obtained from:
dt
2et + a
=
d
4e
(6.56)
=
1
(2et a)
2
(6.57)
for simplicity we adopt
=
0
t
2
(6.58)
Once we have obtained we can obtain f (the scale factor) from
= A
f
3(+1)
= f = (A
t)
2
3(+1)
, (6.59)
in this way we nd H and from eq. (6.86) it is obtained the behaviour of , being this:
c
2
= 3H
2
8G
c
2
(6.60)
therefore
=
_
3
2
8B
0
_
1
c
2
t
2
=
l
c
2
t
2
(6.61)
if we replace all these results into eq. (6.88) then we shall obtain the exact behaviour for c, i.e.
_
1
t
+
c
c
_
=
c
c
(6.62)
being =
l
8B
0
, obtaining in this way:
c = c
0
t
, (6.63)
with =
_

1+
_
.
If we follow each step then it will be obtained the complete solution for each case.
2. Case II, b = e = 0 which imply (t, ) = a, (t, ) = d and therefore:
G
G
= 2
c
c

d
a
(6.64)
that brings us to:
G
c
2
= Kexp(t) (6.65)
where
d
a
=
dt
(t, )
=
d
(t, )
=
dt
a
=
d
d
= =
0
exp(t) (6.66)
this expression only has sense if R
.
The scale factor f veries the relationship:
= A
f
3(+1)
= f = K
f
(exp(t))
1
3(+1)
(6.67)
in this way we nd that
H =

3 ( + 1)
(6.68)
the cosmological constant is obtained from:
c
2
=

2
3 ( + 1)
2
8K
0
= c
2
= l (6.69)
if we replace all these results into eq. (6.88) then we shall obtain the exact behaviour for c, i.e.
_
l
8K
0
+ 2
_
c
c
= (6.70)
nding that
c = Kexp(c
0
t) (6.71)
being c
0
=

_
l
8K
0
+2
_
, note that R
that is to say, c is a decreasing function on time t.

3. Case III.
d4e
2e
= ,
b
2e
= ,
a
2e
=
G
G
2
c
c
=
t +
t +
(6.72)
G
c
2
= K(t + )
exp(t) (6.73)
dt
(t, )
=
d
(t, )
=
dt
t +
=
d
(t + + 2)
= (6.74)
=
0
(t + )
+2
exp(t) (6.75)
following the above steps we nd that the scale factor veries the next relationship
= A
f
3(+1)
= f = K
f
_
(t + )
+2
exp(t)
_ 1
3(+1)
(6.76)
f = K
f
(t + )
1
3
+2
+1
e
1
3(+1)
t
(6.77)
rewriting
+2
3(+1)
= and

3(+1)
=
f = K
f
(t + )
e
t
(6.78)
it is easy nd that
H =

(t + )
+ (6.79)
therefore
c
2
= 3
_
+ (t + )
(t + )
_
2
8
K
0
(t + )
2
= (6.80)
=
K
1
c
2
(t + )
2
+
K
2
c
2
(t + )
+
K
3
c
2
(6.81)
where K
1
=
_
3
2
8K
0
_
, K
2
= 6 and K
3
= 3
2
.
= 2K
1
_
c
c
3
(t + )
2
+
1
c
2
(t + )
3
_
K
2
_
2
c
c
3
(t + )
+
1
c
2
(t + )
2
_
2K
3
c
c
3
(6.82)
In this way we nd that c veries the next eq.
c
c
_
2

K
1
+ 2 +

K
2
(t + ) + 2

K
3
(t + )
2
_
+
2

K
1
+ t +
(t + )
+

K
2
= 0 (6.83)
being

K
i
=
K
i
8K
0
, therefore
ln c =
_

2

K
1
+ t + +

K
2
(t + )
(t + )
_
2

K
1
+ 2 +

K
2
(t + ) + 2

K
3
(t + )
2
_ (6.84)
6.1.3 Fluidos viscosos
The eld equations are:
2
f
f
+
_
f
f
_
2
=
8G
c
2
(p +) +c
2
(6.85)
3
_
f
f
_
2
=
8G
c
2
+c
2
(6.86)
+ 3 ( + 1) H = 3H (6.87)
c
4
8G
+
G
G
4
c
c
= 0 (6.88)
+

k
1
= 3H
1
2
_
3H + W
_
(6.89)
from (6.85) (6.86) it is obtained
2
f
f
2
_
f
f
_
2
=
8G
c
2
(p ++ )
it is observed that
2
f
f
2
_
f
f
_
2
= 2 (H)
then
2 (H)
=
8G
c
2
(p ++ )
from (6.87) we can obtain
H =

3 (( + 1) +)
if we make the assumption = with R
(that may be unfounded, a strong simplication), then we

obtain
H =

3 ( + 1 +)
but in this new approach we are going to study the eq. (6.89) under the assumption = with R
. In
this way we nd that
+

k
1
= 3H
1
2
_
3H +W
+k
1

2
= 3H
1
2
_
3H +W
_
then
H =
_
D
+ k
1

1
_
where =
2
6 + 3
, D =
_
1 +
W
2
_
now we go next to calculate H
,
H
=
_
D
D
_
_
2
+ k
1
(1 )
_
therefore
2 (H)
=
8G
c
2
(p ++ )
(H)
=
4
c
2
G
where = ( + 1 +). Simplifying it yields
=

2
D
1
k
1
+ A
G
c
2
2
(6.90)
with (1 ) = and A =
4
.
As we can see this equation is similar to the eq. under study i.e.
y
= f (x, y, y
) i.e. f (t, ,
) =

2
k
1
+ A
G
c
2
2
Model 1. Constants constants.
X = (t, )
t
+ (t, )
(6.91)
f
t
f
+
tt
+
_
2
t
tt
_
+
_
2
t
_
3
+ ...
... +
_
2
t
3
_
f
_
t
+
_
t
_
_
f
= 0 (6.92)
Simplifying it:
+
1
= 0 (6.93)
1
+
_
2
t
_
+ 2
= 0 (6.94)
_
2
t
tt
_
+
t
K
A
G
c
2
t
2
1
+ K
1
= 0 (6.95)
2A
G
c
2
+
tt
+
A
G
c
2
2
2
t
A
G
c
2
2
+
t
K
= 0 (6.96)
where K = k
1
, note that
f =

2
k
1
+ AG
2
=

2
+ AG
2
f
t
= 0
f
2
k
1
+ 2A
G
c
2
=
2
K
1
+ 2A
G
c
2
= 2
k
1
= 2
We take into account the following dimensional considerations:

_
_
=

,
_
t
_
=

t
= t =
and
_
AG
2
_
= AG,
_
t
AG
2
_
=
AG
2
t
_
t
K
_
=

t
2
we can see that
AG =
AG
2
t
t =
AG =

t
2
AGt
2
= 1
AG
2
t
=

t
2
AG
2
t = t =
therefore with this constrain t = , we solve eqs. (6.93-6.96), nding
(, t) = at + b
(, t) = a
with the help of eq. (6.96)
2Aa
G
c
2
2
+ Aa
G
c
2
2
+ 2aA
G
c
2
2
= 0
we nd that
1 + 2 = 0 =
1
2
(6.97)
where = (1 ) , that is to say =
1
2
.
(, t) =
a
2
t + b, (, t) = a (6.98)
and therefore
X =
_
a
2
t + b
_
t
+ (a)
(6.99)
As we can see, I would like to emphasize the result (6.97) since I believe that it is very important. In this case
Lie method indicates us that there is only one symmetry but iff =
1
2
i.e. the crucial valor which makes scale
invariant the equations etc...
Cases.
1. a = 0 For this value of a eq. (6.90) admits a single symmetry
X =
_
a
2
t b
_
t
(a)
the knowledge of one symmetry X might suggest the form of a particular solution as an invariant of the
dt
(x, y)
=
d
(x, y)
(6.100)
this particular solution is known as an invariant solution (generalization of similarity solution). In this
case
=
0
t
2
for simplicity we have take b = 0, therefore the condition = are very restrictive etc....
Model 2. G and variables.
X = (t, )
t
+ (t, )
(6.101)
f
t
f
+
tt
+
_
2
t
tt
_
+
_
2
t
_
3
+ ...
... +
_
2
t
3
_
f
_
t
+
_
t
_
_
f
= 0 (6.102)
Therefore the new results are:
+
1
= 0 (6.103)
1
+
_
2
t
_
+ 2
= 0 (6.104)
_
2
t
tt
_
+
t
K
A
G
c
2
t
2
1
+ K
1
= 0 (6.105)
A
G
c
2

2
2A
G
c
2
+
tt
+
A
G
c
2
2
2
t
A
G
c
2
2
+
t
K
= 0 (6.106)
where K = k
1
, note that
f =

2
+ A
G
c
2
2
f
t
= A
G
c
2

2
f
2
K
1
+ 2A
G
c
2
= 2
therefore with this constrain t = , we solve eqs. (6.103-6.106), nding

(, t) = at + b, (, t) = a
with
G
G
=
(1 2)a
at b
where = (1 ) .
Cases We have only two simple cases:
1. a = 0 which correspond to G = const.
2. a = 0 which correspond to
G = G
(t + b)
2+
1
it is observed that if =
1
2
then G = const. since = (1 ), if >
1
2
then G is a growing function on
time and if <
1
2
then G is a decreasing function on time. For this form of G eq. (6.90) admits a single
symmetry
X = (at b)
t
(a)
the knowledge of one symmetry X might suggest the form of a particular solution as an invariant of the
dt
(x, y)
=
d
(x, y)
(6.107)
this particular solution is known as an invariant solution (generalization of similarity solution). In this
case
t
=
for simplicity we have take b = 0.
With this result we can apply the standard Lie procedure or through pedestrian method to try to obtain
the rest of quantities like . In this way taking into account dimensional considerations from eq.
=

2
k
1
+ AG
2
it is obtained the next relationship between quantities
k
1
t 1 and AGt
2
1
this last relationship is the known relation for the inertia obtained by Sciama. From these relationship we
obtain
A
1
(t + b)
2
1
t
2
A
1
(t)
we can see that this result veries the relation

k
1
t 1 since k
1
t
1
t 1
therefore the condition = are very restrictive.
Model 3. G, c and variables.
X = (t, )
t
+ (t, )
(6.108)
f
t
f
+
tt
+
_
2
t
tt
_
+
_
2
t
_
3
+ ...
... +
_
2
t
3
_
f
_
t
+
_
t
_
_
f
= 0 (6.109)
+
1
= 0, (6.110)
1
+
2
t
+ 2A
1s
= 0, (6.111)
2
t
tt
+ A
t
1s
3B
G
c
2
2
2
t
1
+ A(1 s)
s
= 0, (6.112)
B
_

G
c
2
2G
c
c
3
_
2
2B
G
c
2
+
tt
+
_
2
t
_
B
G
c
2
2
+ A
t
1s
= 0, (6.113)
(, t) = ast + b, (, t) = a (6.114)
with the constraint
G
G
= 2
c
c
+
(1 2s)a
b ast
, (6.115)
being a and b numerical constant.
Thus we have found all the possible forms of G and c for which eq. (6.90) admits symmetries. There are two
cases with respect to the values of the constant a, a = 0 which correspond to G = const, c = const. and a = 0
which correspond to
G
c
2
= K [ast + b]
2+
1
s
, (6.116)
with K a constant of integration. If s =
1
2
, which correspond to =
1
2
, then G and c remain constants. For this
form of G and c eq. (6.90) admits a single symmetry
X = (ast b)
t
(a)
. (6.117)
this is not a surprising result since as already have been point out by M. Szydlowki and M. Heller there is only
one symmetry.
The knowledge of one symmetry X might suggest the formof a particular solution as an invariant of the operator
X, i.e. the solution of
dt
(t, )
=
d
(t, )
(6.118)
This particular solution is known as an invariant solution (generalization of similarity solution). In this case
=
0
t
1
s
, (6.119)
where for simplicity we have taken b = 0, and
0
is a constant of integration (note that s = (1 ) being the
bulk viscous parameter).
We can apply a pedestrian method to try to obtain the same results. In this way, taking into account dimensional
considerations, from the eq. (6.90) we obtain the following relationships between density, time and gravitational
constant:
k
1
s
s
t 1, B
G
c
2
t
2
1. (6.120)
This last relationship is also known as the relation for inertia obtained by Sciama. From these relationship we
obtain
B
1
K
1
c
2
[ast + b]
2s1
s
t
2
B
1
K
1
c
2
[ast]
1
s
. (6.121)
We can see that this result veries the relation
k
1
s
s
t 1. (6.122)
Once we have obtained we can obtain f (the scale factor) from
= A
,
f
3(+1+)
= f =
_
A
,
1
0
t
_ 1
3(+1+)s
, (6.123)
in this way we nd H and from eq. (6.86) it is obtained the behaviour of , being this:
=
_
3
2
8K(as)
2+
1
s

0
_
c
2
t
2
=
l
c
2
t
2
(6.124)
being =
(A
,
1
0
)
3(+1+)s
if we replace all these results into eq. (6.88) then we shall obtain the exact behaviour for c,
i.e.
_
l
4
K
0
+
1 2s
s
_
1
t
=
_
l
4
K
0
+ 2
_
c
c
, (6.125)
therefore
c = c
0
t
, (6.126)
being =
_
+
(12s)
s
_
(+2)
and =
l
4
K
0
.
Therefore the conditions = together with the equation (6.89) are very restrictive. Hence we have shown
that under these assumptions there are no another possible solutions for the eld equations.

Ecuaciones Diferenciales Mediante Simetrías

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Ecuaciones Diferenciales Mediante Simetrías

Hochgeladen von

Copyright:

Verfügbare Formate

Ecuaciones Diferenciales

are measured in term of m fundamental dimensions labeled as dimensional

, in this way eq. (1.1) is equiv-

satisfy assumptions (iii), and

= (x, y), (2.8)

= (x, y). (2.9)

= (x, y), (2.10)

) the linearized symmetry condition is obtained substituting (2.36) and

, Equation (2.45) can be separated with respect to powers of y

we nd the following symmetry

g(y) (x, y) f (x)g

que podemos sustituir en (3.93),

que hemos obtenido

esta ecuacin es d.h. sii [a] = x

66 CHAPTER 3. ECUACIONES DE PRIMER ORDEN

= t(u + 2) + 2u, (3.242)

= [b] = [e] , having only two dimensional constants. Since y = x

cuya solucin es:

= d f /dx, y cuya solucin de (4.55) es:

= p(t)u + q(t) (4.134)

we obtain the overdetermined system:

that is to say, c is a decreasing function on time t.

(that may be unfounded, a strong simplication), then we

We take into account the following dimensional considerations:

therefore with this constrain t = , we solve eqs. (6.103-6.106), nding

we can see that this result veries the relation

Das könnte Ihnen auch gefallen