International Journal of Contemporary Mathematical Sciences

Vol. 10, 2015, no. 7, 311 - 316

HIKARI Ltd, www.m-hikari.com
http://dx.doi.org/10.12988/ijcms.2015.5732

On the Limit Cycles for Li

enard Equation
Osuna Osvaldo, Rodrguez Joel, Villase
nor Gabriel
and Villavicencio Geiser
Instituto de Fsica y Matematicas, Universidad Michoacana
Edif. C-3, Cd. Universitaria
C.P. 58040. Morelia, Mich., Mexico
Instituto Tecnologico de Morelia
Departamento de Ciencias Basicas, Edif. AD
Morelia Michoacan, Mexico
Depto. de Ciencias Ambientales, UAM-Lerma
Av. Hidalgo Pte. No. 46, col. La Estacion
Lerma de Villada, Edo. de Mexico, Mexico
c 2015 Osuna Osvaldo et al. This article is distributed under the Creative
Copyright
Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract
In this work, we present some new criteria on the non-existence and
uniqueness of limit cycles for the Lienard equation.

Mathematics Subject Classification: 34C05, 34C25

Keywords: Dulac functions, limit cycles, Lienard equation

Introduction

We consider the Lienard equation given by


x 1 = x2 ,
x 2 = g(x1 ) f (x1 )x2 ,

(1)

312

Osuna Osvaldo et al.

It is well-known the relevance of the Lienard equation in the qualitative theory

of differential equations, which models several oscillatory phenomena. We are
concerning on the non-existence and the uniqueness of limit cycles, for this we
use of the extended Bendixson-Dulac criterion (see [1] and [2]).
Given an open set R2 we consider


x 1 = f1 (x1 , x2 ),
x 2 = f2 (x1 , x2 ), (x1 , x2 ) ,

(2)

where f1 , f2 are C 1 -functions on , the associated vector field is F (x1 , x2 ) =

f2
f1
+ x
.
(f1 (x1 , x2 ), f2 (x1 , x2 )). As usual its divergence is div(F ) = x
1
2
We consider the sets
F := {f C 0 (, R {0}) : vanishes only on a measure zero set},
and F := F F+ , along the paper we will use the Lebesgue measure.
Recall that an open subset R2 intuitively is said to be l-connected if it
has l-holes, i.e., if its first fundamental group is a free group with l-generators,
we denote l() = l.
For h : R a continuous function, let Z(h) := {x : h(x) = 0} be
the set of zeros of h.
Following ([2]) we denote by l(, h) the sum of the quantities l(U ) over all
the connected components U of \ Z(h), also denote by co(h) the numbers of
closed ovals of Z(h) contained in .
Our results are established with the help of the techniques developed in [4],
let us recall the following result
Proposition 1.1 ([4], prop. 2) Let R2 be an open set with regular
boundary. Suppose that there is s R and a function c : R such that
hh, F i + shdiv(F ) = ch,

(3)

admits an analytic solution h, with ch does not change sign and vanishes only
on a null measure subset, then h is a Dulac function. Therefore the limit cycles
of system (2) are either totally contained in Z(h), or do not intersect Z(h).
Moreover, the number of limit cycles contained in Z(h) is at most co(h) and
the number N of limit cycles that do not intersect Z(h) satisfies

if s > 0,
l()
0
if s = 0,
N

l(, h) if s < 0.

(4)

313

Limit cycles for Lienard equation

Results and applications

Our first result gives a criterion for the non existence of limit cycles for (1)
Proposition 2.1 If there are c1 , c2 constants, such that c1 g(x1 )+ c2 f (x1 )
FR2 , then the Lienard system has not periodic orbits.
Proof: Seeking for a function of Dulac h = h(z), depending only on z =
z(x1 , x2 ), the associated equation (3) becomes


z d ln h
z
+ (g(x1 ) f (x1 )x2 )
= c + sf (x1 ),
x2
x1
x2
dz
taking z such that
and

Rx
z
z
f (x1 )
= 0; hence we obtain z = 1 f (s)ds + x2
x1
x2
c + sf (x1 )
d ln h
=
,
dz
g(x1 )

(5)

where the right side depends on z, denoted this by (z); so taking

c = c1 f (x1 ) + c2 g(x1 ) and s = c1 ,
d ln h
= c2 , whose solution
then (z) = c2 and the equation (5) is written as
dz
is
Z x
1

h(x1 , x2 ) = ez = exp[c2 (

f (s)ds + x2 )].

Note that Z(h) = contains no ovals. In particular, co(h) = 0.

Since l(R2 , h) = l(R2 ) = 0, then by Proposition 1.1, the equation (1) has
no limit cycles, so the result follows.
2
Corollary 2.2 If any of the following conditions holds
1. g(x1 ) or f (x1 ) belong to FR2 ,
2. kf (x1 ) > g(x1 ), for some k constant,
then the Lienard system has not periodic orbits.
Remark: The condition 1 is a slight generalization of Criterion 1, (C1) in [3].
Theorem 2.3 Assume g polynomial.R If for some a R is satisfied that
x
f (x1 )(G(x1 ) + a) FR2 where G(x1 ) = 0 1 g(s)ds, then the Lienard equation
has at most one limit cycle.

314

Osuna Osvaldo et al.

Proof: Assuming that h = h(z), depends on z = z(x1 , x2 ), the associated

equation (3) is written as


z d ln h
z
(g(x1 ) + f (x1 )x2 )
= c + sf (x1 ),
(6)
x2
x1
x2
dz
z
z
g(x1 )
= 0; hence we get z =
x1
x2
for some constant a and (6) becomes

taking z such that x2

d ln h
c + sf (x1 )
=
,
dz
f (x1 )x22

x22
2

R x1

g(s)ds + a

(7)

where the right side depends on z, denoted this by , thus c = (z) (f (x1R)x22 )
2 x1 g(s)ds+a
sf (x1 ), we choose s = 2 and (z) = z1 ; simplifying we have c = f (x1 )
,
z
ln h
ln z
taking c in this way, the equation (7) is written as d dz
= d dz
whose solution
is
Z x1
x22
+
g(s)ds + a.
h(x1 , x2 ) = z =
2
Rx
By hypothesis the function ch = 2f (x1 )( 1 g(s)ds + a) is continuous does
not change sign and vanishes only on a measure zero set.
Note that Z(h) has at most one oval, and therefore co(h) = 1; nevertheless,
Z(h) contains no limit cycles, because the vector field is not tangent to z = 0.
Since s < 0 and l(R2 , h) is less or equal than 1, therefore by Proposition
1.1 the system (1) has at most one limit cycle.
2
Observe that Theorem 2.3 recovers a result due to Sansone [5]
Corollary 2.4 Sansones theorem: If g(x1 ) = x1 , and there exists > 0
such
R that f (x1 ) < 0 for x1 (, ), f (x1 ) > 0 in (, ) (, +) also
f (s)ds = , then system (1) has at most one limit cycle.
0
Rx
2
x2
Proof: Indeed, taking G(x1 ) = 1 g(s)ds = 21 and a = 2 , the hypotheses
of Theorem 2.3 are fulfilled, therefore (1) has at most one limit cycle.
2
Proposition 2.5
R x1If there is m R such that f (x1 )[2F (x1 ) + m] g(x1 )
FR2 with F (x1 ) = 0 f (s)ds, then the Lienard equation has not limit cycles.
z
z
f (x1 )
=
Proof: Consider the equation (6) and taking z such that
x1
x2
R x1
f (x1 ); hence we get z = x2 + 2
f (s)ds + a for some constant a and (6)
becomes
d ln h
c + sf (x1 )
=
,
(8)
dz
[x2 f (x1 ) g(x1 )]

315

Limit cycles for Lienard equation

where the right side depends on z, we take this equal to

c=

1
; so
z

x2 f (x1 ) g(x1 ) + zsf (x1 )

z

then, replacing z in the numerator and choosing s = 1, a = m, we get

equation (5) is written as
c = g(x1 )+f (x1z)[2F (x1 )+m] , taking c in this way,
R xthe
1
d ln h
1
=
,
whose
solution
is
h(x
,
x
)
=
z
=
2
f
(s)ds
+ x2 m.
1
2
dz
z
2
Note that Z(h) is a submanifold of R homeomorphic to R therefore no
limit cycles are contained in Z(h).
Since l(R2 , h) = 0, then by Proposition 1.1, (1) has not limit cycles.
2
Proposition 2.6 If there exist a function (x1 ) strictly monotonic such
that it satisfies the inequality
2 (x1 )f 2 (x1 ) + 4 0 (x1 )(x1 )g(x1 ) FR2 ,
then the Lienard system has not periodic orbits.
Proof: Suppose the function h has the form h = (x1 )x2 , then the associated
equation becomes
x22

d(x1 )
(g(x1 ) + f (x1 )x2 )(x1 ) = h[c + sf (x1 )].
dx1

Taking s = 0 and as by conditions of Proposition 1.1 ch must be in FR2 , we

need that
x22 x1 (x1 ) (x1 )f (x1 )x2 (x1 )g(x1 ) FR2 .
Now this inequality holds if 0 0, ( 0 0) and the discriminant
2 (x1 )f 2 (x1 ) + 4x1 (x1 )g(x1 ) FR2 .
But it holds by hypothesis.

Corollary 2.7 If g(x1 ) is strictly decreasing and satisfies f 2 (x1 ) + 4gx1

FR2 , then the Lienard system has not periodic orbits.

References
[1] L. Cherkas, Dulac functions for polynomial autonomous systems on a
plane, Diff. Eqs., 33 (1997), 692 - 701.
[2] A. Gasull, H. Giacomini, Some applications of the extended BendixsonDulac theorem, Progress and Challenges in Dyn. Syst., 54 (2013), 233 252. http://dx.doi.org/10.1007/978-3-642-38830-9 14

316

Osuna Osvaldo et al.

[3] A. Gasull, H. Giacomini, J. Llibre, New criteria for the existence

and non-existence of limit cycles in Lienard differential systems, Dynamical Systems: An Int. Journal, 24 (2009), no. 2, 171 - 185.
http://dx.doi.org/10.1080/14689360802568086
[4] O. Osuna, G. Villase
nor-Aguilar, On the Dulac functions for multiply
connected domains, E. J. of Qual. Th. of Diff. Eqs., 61, (2013), 1 - 11.
http://dx.doi.org/10.14232/ejqtde.2013.1.61
[5] G. Sansone, Sopra lequazione di A. Lienard delle oscillazioni di rilassamento, Ann. Math. Pura Appl., Series 4, 28 (1949), 153 - 181.
http://dx.doi.org/10.1007/bf02411124
Received: July 27, 2015; Published: August 25, 2015