Skip to main content
QUICK REVIEW

[Paper Review] Remarks on iterated cubic maps

John Milnor|ArXiv.org|May 12, 1990
Mathematical Dynamics and FractalsMathematics30 references124 citations
TL;DR

This paper investigates the dynamics of iterated cubic maps on the real and complex lines, focusing on the parameter space structure and classification of real cubic polynomials via topological conjugacy classes. It introduces a moduli space using invariants A = a² and B = b², with real maps further classified by sign(σ) = sign(B) when B ≠ 0, and by σ = ±1 when B = 0, leading to a disjoint union of half-planes. The key contribution is a classification of real cubic maps into four dynamical classes (ℛ₀ to ℛ₃), with ℛ₃ maps exhibiting Cantor set Julia sets and maximal topological entropy log(3).

ABSTRACT

This note will discuss the dynamics of iterated cubic maps from the real or complex line to itself, and will describe the geography of the parameter space for such maps. It is a rough survey with few precise statements or proofs, and depends strongly on work by Douady, Hubbard, Branner and Rees.

Motivation & Objective

  • To understand the global structure of the parameter space for cubic polynomial maps under affine conjugation.
  • To classify real cubic maps into dynamical classes based on the topology of their real filled Julia sets and critical orbit behavior.
  • To clarify the role of the sign of the third derivative (σ) as an invariant in the real case, especially when B = 0.
  • To describe the geometry of hyperbolic components and boundary structures in the (A,B)-parameter plane.
  • To establish connections between real dynamics, complex dynamics, and the structure of Hubbard trees for cubic maps.

Proposed method

  • Uses affine conjugation to reduce any cubic polynomial to the normal form f(z) = z³ - 3a²z + b, with critical points at ±a.
  • Defines moduli space coordinates A = a² and B = b², which classify complex cubic maps up to affine conjugation.
  • Introduces the real invariant σ = sign(g′′′) to distinguish real affine conjugacy classes when B = 0.
  • Classifies real cubic maps into ℛ₀ to ℛ₃ based on the number of connected components in the graph of f over the minimal invariant interval I.
  • Applies techniques from complex dynamics, including iteration of critical points and estimation of escape time using partial derivatives.
  • Employs numerical algorithms to compute parameter plane images, detecting hyperbolic components and boundaries via periodicity detection and derivative blowup.

Experimental results

Research questions

  • RQ1How is the moduli space of complex cubic maps parameterized, and what are the invariants A and B?
  • RQ2What additional real conjugacy invariant arises when B = 0, and how does it affect the classification of real cubic maps?
  • RQ3How do the dynamical classes ℛ₀ to ℛ₃ correspond to the number of components in the graph of f over the invariant interval I?
  • RQ4What dynamical properties characterize maps in the class ℛ₃, particularly in terms of Julia sets and topological entropy?
  • RQ5How do the boundaries of hyperbolic components in the (A,B)-plane reflect changes in periodic orbit structure and critical orbit behavior?

Key findings

  • The moduli space of complex cubic maps is parameterized by A = a² and B = b², with maps conjugate if and only if (A,B) are equal.
  • For real cubic maps, the invariant σ = sign(g′′′) distinguishes two real affine conjugacy classes when B = 0, even though the complex conjugacy class is the same.
  • Maps in the class ℛ₃ (d = 3) have all critical orbits escaping to infinity, and their real filled Julia set Kℝ is a Cantor set of measure zero.
  • The restriction f|Kℝ is topologically conjugate to a one-sided shift on three symbols, giving maximal topological entropy log(3).
  • For maps in ℛ₃, the complex Julia set coincides with the real Cantor set Kℝ, implying all complex periodic points are real and in Kℝ.
  • The boundary of hyperbolic components in the (A,B)-plane is detected by large partial derivatives of iterates with respect to A and B, or by slow convergence to periodic orbits near parabolic points.

Better researchstarts right now

From reading papers to final review, dramatically reduce your research time.

No credit card · Free plan available

This review was created by AI and reviewed by human editors.