[Paper Review] Dominated chain recurrent class with singularities

Christian Bonatti, Shaobo Gan+1 more|arXiv (Cornell University)|Jun 20, 2011

Mathematical Dynamics and Fractals28 references3 citations

TL;DR

This paper proves that for generic three-dimensional vector fields, any chain recurrent class exhibiting a dominated splitting must be singular hyperbolic, provided it contains singularities. The result establishes that domination in three-dimensional flows implies stronger hyperbolic-like behavior—singular hyperbolicity—thereby supporting Palis' conjecture that robustly non-hyperbolic dynamics are structurally approximated by systems with singular cycles or Lorenz-like attractors.

ABSTRACT

We prove that for generic three-dimensional vector fields, domination implies singular hyperbolicity.

Motivation & Objective

• To understand the global dynamics of generic three-dimensional vector fields, particularly in the presence of singularities.
• To investigate whether domination in a chain recurrent class implies singular hyperbolicity, a generalization of hyperbolicity for systems with singularities.
• To verify a key conjecture in dynamical systems: that robustly non-hyperbolic dynamics in 3D are approximated by systems with singular cycles or Lorenz-like attractors.
• To clarify the structure of chain recurrent classes in the presence of singularities under generic $C^1$ conditions.
• To establish that non-trivial chain recurrent classes with domination and singularities must be singular hyperbolic, thus extending hyperbolic theory to non-hyperbolic settings.

Proposed method

• Uses the concept of dominated splitting: a continuous, invariant splitting $T_{\Lambda}M = E \oplus F$ such that $\|\Phi_t|_E\| \|\Phi_{-t}|_F\| \leq Ce^{-\lambda t}$ for some $C>0$, $\lambda>0$.
• Applies the notion of singular hyperbolicity: a partially hyperbolic splitting where $E$ is sectional contracting and $F$ is sectional expanding (or $E$ contracting and $F$ sectional expanding).
• Employs $C^1$ genericity assumptions to ensure that unstable manifolds of singularities are dense in the chain recurrent class.
• Uses Lyapunov stability and chain recurrence properties to analyze the omega-limit sets of regular points and their accumulation toward singularities.
• Applies compactness arguments and estimates on the norm of the tangent flow $\Phi_t$ to prove uniform contraction on the $E$ bundle.
• Utilizes the $C^1$ connecting lemma and genericity to show that sinks accumulating toward a chain recurrent class must be contained in it, leading to contradiction if the class is not singular hyperbolic.

Experimental results

Research questions

• RQ1 Does domination in a chain recurrent class of a generic 3D vector field imply singular hyperbolicity?
• RQ2 Can a non-trivial chain recurrent class with singularities and a dominated splitting fail to be singular hyperbolic?
• RQ3 What is the role of singularities of index 1 and 2 in the structure of chain recurrent classes under domination?
• RQ4 How do sinks and periodic orbits accumulate in a chain recurrent class, and what constraints does this place on the dynamics?
• RQ5 To what extent do generic $C^1$ vector fields on 3-manifolds satisfy Palis' conjecture on robust non-hyperbolic dynamics?

Key findings

• For generic $C^1$ vector fields on a 3-manifold, any chain recurrent class with a dominated splitting is singular hyperbolic if it contains singularities.
• If a chain recurrent class $C(\sigma)$ contains a singularity $\sigma$ of index 2, then the dominated splitting $E \oplus F$ satisfies $\dim E = 1$ and $E$ is uniformly contracting.
• The class $C(\sigma)$ is Lyapunov stable, and every singularity in $C(\sigma)$ has index 2, implying no regular points can accumulate toward the stable manifold of a singularity of index 1.
• The existence of a sequence of sinks accumulating toward $C(\sigma)$ leads to a contradiction unless $C(\sigma)$ is singular hyperbolic, thus proving the necessity of singular hyperbolicity under domination.
• If $C(\sigma)$ contains a singularity and is not singular hyperbolic, then it must contain a hyperbolic periodic orbit whose unstable manifold is dense in $C(\sigma)$, leading to contradiction via sink accumulation.
• The result confirms that singular hyperbolicity is the generic outcome of domination in 3D flows with singularities, supporting the broader conjectural framework of Palis.