Bifurcations of limit cycles in a cubic system with cubic perturbations☆
Introduction
There have been many studies on the number and distributions of limit cycles for planar polynomial systems (see [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18] for example). There are three main aspects to study the number of limit cycles for perturbations of Hamiltonian systems as follows. These are:
(1) Hopf bifurcation. The main idea is to compute Liapunov constants or focus values for a focus and then change the stability of the focus by varying focus values to obtain limit cycles one by one (see [1], [9], [14], [15]). The maximal number of limit cycles which can appear in a neighborhood of the focus under perturbations is called the Hopf cyclicity of the focus about the considered system. The Hopf cyclicity can also be obtained by coefficients appearing in the expression of the first Melnikov function near a center (see [3]).
(2) Poincare bifurcation. The main task in Poincare bifurcation is to find the maximal number of limit cycles generated by members of a family of periodic orbits. There have been a lot of studies on this aspect (see [9], [10], [11], [12], [14], [15]). The number can be obtained through the study of the total number of zeros for the first or higher order Melnikov functions using the method of geometric analysis. Using this method eleven limit cycles were obtained in [10], [11], [12] for the following cubic systems:with a > c > 0, 0 < μ ≪ 1, andwith a > c > 0, ac > 1, 0 < ε ≪ 1.
(3) Homoclinic or heteroclinic bifurcation. In 1986, Roussarie [13] originated a way to study the number of limit cycles appearing near a homoclinic loop and obtained a significant result to find the homoclinic cyclicity by using coefficients in the expression of the first order Melnikov function. Then the idea was developed and more results were obtained in [2], [4], [5], [6]. It consists of the following three steps:
- •
find discriminate values to determine the stability of the homoclinic or a double homoclinic loop;
- •
vary parameters to change the stability of the loop to produce limit cycles;
- •
find a final limit cycle by breaking the homoclinic loop.
Recently, eleven limit cycles were found near double homoclinic loops by using this method in [8], [16] and a new distribution of 11 limit cycles was found. The systems studies in [8], [16] are, respectively,with c > 1, 0 < μ ≪ 1, andwith 0 > c > −1 > a, ac > 1.
We consider the following perturbed cubic Hamiltonian systemwhere a and c are positive constants and ε > 0 is small. Without loss of generality, we may assume a ⩾ c > 0, ac > 1. Also, we consider the coefficients aij and bij in (1.1) as parameters. Our main result can be stated as the conclusion in Section 5 and the distributions of limit cycles see Fig. 1.1.
In Section 2, we give the phase portraits of the unperturbed system. In Section 3, we verify the results following the technique developed in [2], [5], [8], [16] and consider the case of a ≠ c and a = c separately. For convenience of numerical analysis we will suppose a = 2 and c = 1 for the first case and a = c = 2 for the second.
Section snippets
The phase portraits of the unperturbed systems
In this section we study the phase portraits of the unperturbed systems.
The proof of the conclusion: Case 1
Consider (1.1). Without loss of generality, we suppose the points , are always the saddle points of system (1.1) (if necessary making a linear transformation of variables). Then (1.1) becomeswhere 0 < ε ≪ 1. Since system (3.1) is symmetric with respect to the origin, we only consider the half-plane of x ⩾ 0 in the following. Let
The proof of the conclusion: Case 2
In this section we consider the following perturbed systemwhere a > 1, 0 < ε ≪ 1. Since (4.1) is symmetric with respect to the origin, as before we only consider the half-plane of x ⩾ 0 in the following. Let , , , ω(L1) = S3, ω(L2) = S1, ω(L3) = S4, ω(L4) = S1. For ε > 0 small, the (4.1) has separatrice
Conclusion
Let a ⩾ c > 0, ac > 1. Then the system (1.1) can have 11 limit cycles. The distributions of limit cycles are given in the Fig. 1.1.
References (18)
On the Hopf Cyclicity of plane systems
J. Math. Anal. Appl.
(2000)- et al.
On the stability of double homoclinic and heteroclinic cycles
Nonlinear Anal.
(2003) - et al.
Bifurcations of limit cycles in a cubic system
Chaos Soliton Fract
(2004) - et al.
Bifurcations of limit cycles for a cubic Hamiltonian system under quartic perturbations
Chaos Soliton Fract
(2004) - et al.
Bifurcations of limit cycles from quintic Hamiltonian systems with a double figure eight loop
J. Complex.
(2004) On the number of limit cycles which appear with the variation of coefficients from an equilibrium position of focus or center type
Mat. Sbornik (N.S.)
(1952)Cyclicity of planar homoclinic loops and quadratic integrable systems
Sci. China, Ser. A
(1997)- et al.
Cyclicity 1 and 2 conditions for a 2-polycycle of integrable systems on the plane
J. Different. Equat.
(1999) - et al.
On the number of limit cycles in double homoclinic bifurcations
Sci. China, Ser A
(2000)
Cited by (4)
Investigation of relation between singular points and number of limit cycles for a rotor-AMBs system
2009, Chaos, Solitons and FractalsThirteen limit cycles for a class of cubic Hamiltonian system with higher-order perturbed terms
2008, Applied Mathematics and ComputationDetection function method and its application to a class of quintic Hamiltonian systems with quintic perturbations
2007, Applied Mathematics and ComputationOn the number of limit cycles of a cubic polynomials Hamiltonian system under quintic perturbation
2007, Applied Mathematics and Computation
- ☆
Supported by Shanghai Leading Academic Discipline Project (No. T0401) and Australian Research Council (ARC) under the Discovery Projects Scheme (Grant ID: DP0559111).