Mingzhi Hu

1 · 2026

Read Mingzhi Hu by 1 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-24612.txt.

Introduction

Hu and co-authors construct a three-dimensional competitive Lotka-Volterra system with four limit cycles. The example belongs to class 28 in Zeeman's classification of such systems. Prior work had already found systems with at least four limit cycles for neighbouring classes. Gyllenberg and Yan handled class 27, Wang, Huang and Wu class 29, and Yu, Han and Xiao class 26. With this new construction, every class from 26 to 29 is now known to admit a system with at least four limit cycles.

Catalog record

CREATOR1
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE18534
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.24612

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Mingzhi Hu 1 Abstract Four Limit Cycles in Three-Dimensional Competitive Lotka-Volterra Systems of Class 28 in Zeeman’s Classification Abstract In this paper, a three-dimensional competitive Lotka-Volterra system with four limit cycles is constructed for class 28 in Zeeman’s classification. Combined with existing results — from Gyllenberg and Yan (2009) for class 27, from Wang, Huang and Wu (2011) for class 29, and from Yu, Han and Xiao (2016) for class 26 — our finding indicates that there exist systems with at least four limit cycles for each class among classes 26 - 29. Keywords: Lotka-Volterra; competitive systems; Zeeman’s classification; class 28; four limit cycles Introduction MacArthur [ 16 ] constructed a Lyapunov function for a Lotka-Volterra system with a symmetric interaction matrix, complementing the original Lyapunov function proposed by Volterra [ 20 ] . By means of MacArthur’s Lyapunov function, it can be shown that an \omega -limit set for a competitive system with a symmetric interaction matrix is either an isolated equilibrium or a continuum of equilibria. In 1975, May and Leonard [ 17 ] presented a landmark result for three-dimensional competitive systems. They constructed a three-dimensional competitive Lotka-Volterra system with a cyclically symmetric interaction matrix and obtained a repulsive heteroclinic cycle.Meanwhile, through numerical simulation, Gilpin [ 3 ] found that a three-dimensional competitive system may admit a limit cycle. In 1979, Coste, Peyraud and Coullet [ 2 ] and Schuster, Sigmund and Wolff [ 18 ] independently proved the existence of a limit cycle for three-dimensional competitive systems (which actually belong to class 27 in Zeeman’s classification [ 24 ] ), and verified the attractiveness or repulsion of the corresponding heteroclinic cycle. In 1981, Hofbauer [ 7 ] proved the existence of limit cycles in general n -dimensional (n\geqslant 4) competitive systems via Hopf bifurcation theory. Smale [ 19 ] in 1976 and Hirsch [ 9 ] in 1988 demonstrated that the dynamical behaviors of n -dimensional competitive systems are equivalent to those of (n-1) -dimensional systems.For a three-dimensional competitive Lotka-Volterra system, Hirsch’s result [ 9 ] ensures that there exists an invariant manifold (called a carrying simplex), which is homeomorphic to a two-dimensional simplex and attracts all orbits except the origin. Based on Hirsch’s theorem, Zeeman [ 24 ] used the qualitative analysis of the system at the boundary surface to define the combinatorial equivalence relation through parameter inequalities, and obtained thirty-three stable classes of three-dimensional Lotka-Volterra competitive systems. Zeeman’s result indicates that the limit sets for 27 classes among the 33 ones are all fixed points [ 24 ] , so the dynamical behaviors of systems in these 27 classes are fully described. Using the Hopf bifurcation theorem, Zeeman further proved that for each of the remaining six classes (class 26 to class 31), there always exists a system that can be constructed to admit an isolated periodic orbit or limit cycles via appropriate parameter selection. The first result of a three-dimensional competitive Lotka-Volterra system with two limit cycles was obtained in class 27 of Zeeman’s classification by Hofbauer and So [ 8 ] based on Hirsch’s monotonic flow theorem, the center manifold theorem and the Hopf bifurcation theorem. In their example, a locally stable positive equilibrium is surrounded by two limit cycles: one is from the Hopf bifurcation theorem, and the other is guaranteed via the Poincaré-Bendixson theorem. In [ 8 ] , Hofbauer and So proposed a problem of how many limit cycles may exist in the six classes 26 - 31 in Zeeman’s classification. In 2002, Lu and Luo [ 14 ] constructed two limit cycles in class 26, class 28, and class 29, respectively, which partially answered Hofbauer and So’s question about the existence of two limit cycles in other classes. In 2003, the first attempt to construct three limit cycles was made by Lu and Luo [ 15 ] , who proposed a system with three limit cycles in Zeeman’s class 27. However, as Yu, Han and Xiao [ 23 ] pointed out, due to a sign error, the positive definiteness of the Lyapunov function for calculating the focal values was not guaranteed, and thus Lu and Luo’s original attempt was invalid. In 2006, Gyllenberg, Yan and Wang [ 4 ] found that the system with two limit cycles in class 29 constructed by Lu and Luo [ 14 ] can actually admit a third limit cycle. The instability of the outer small-amplitude limit cycle, together with the repulsion of the boundary of the carrying simplex, satisfies the conditions of the Poincaré-Bendixson theorem, which guarantees the existence of a third limit cycle. There were also works by Gyllenberg and Yan [ 5 ] in 2009 and Wang, Huang and Wu [ 21 ] in 2011, respectively, for the existence of multiple limit cycles for class 30. Gyllenberg and Yan claimed to have constructed two limit cycles in class 30. One is a small-amplitude limit cycle obtained via Hopf bifurcation, and the other is obtained via the Poincaré-Bendixson theorem. As pointed out by Yu, Han and Xiao [ 23 ] , their construction of limit cycles was incorrect since the Lyapunov function they gave is not positive definite. However, with careful calculation, three limit cycles can actually be constructed by using their original example [ 5 ] [ 23 ] . First, two small-amplitude limit cycles can be constructed such that the outer one is stable. Since the system belongs to class 30, the boundary of the carrying simplex is an attractor; therefore, the third limit cycle can be obtained via the Poincaré-Bendixson theorem. Three years later, Wang, Huang, and Wu [ 21 ] proved that for each class in Zeeman’s classes 26 - 31, there exists a system with at least three small-amplitude limit cycles. Unfortunately, their result for class 30 was invalid, as the constructed system was not competitive. In 2009, Gyllenberg and Yan [ 6 ] constructed an example in class 27 similar to [ 15 ] and claimed that they obtained four limit cycles, three of which were obtained due to the Hopf bifurcation, and the fourth one was obtained through the existence of the heteroclinic cycle and using the Poincaré-Bendixson theorem. Similarly, Gyllenberg and Yan neglected the positive definiteness of the Lyapunov function. Yu, Han and Xiao [ 23 ] pointed out that there does exist an example in [ 6 ] belonging to class 27, which can admit four limit cycles. A system in class 29 with four limit cycles was given by Wang, Huang, and Wu [ 21 ] . Recently, Yu, Han, and Xiao [ 23 ] gave two examples in class 27 and two in class 26 with four small-amplitude limit cycles, respectively. Besides, Li and Jiang [ 10 ] showed three limit cycles for classes 28 and 31. In summary, multiple limit cycles can appear in each of the six classes of Zeeman’s classification: classes 26, 27, 28, 29, 30 and 31. The known results to date are as follows: for the classes 26, 27 and 29, there are at least four limit cycles. An open problem proposed by Yu, Han and Xiao [ 23 ] is whether there are four limit cycles in all classes 26—31. In this paper, by combining

Read the complete text