[Paper Review] Global angle-action variables for Duffing system

Isaak A. Kunin, Alexander Runov|ArXiv.org|May 17, 2001

Quantum chaos and dynamical systems4 references3 citations

TL;DR

This paper introduces a global action-angle variable formulation for the Duffing system by leveraging topological covering of the phase space via a double-cover transformation, enabling consistent action-angle dynamics across topologically nontrivial regions. The method achieves global action-angle variables by exploiting symmetry and embedding the system in a branched covering space, allowing for consistent perturbation analysis even beyond simply connected domains.

ABSTRACT

The classical representation of Hamiltonian systems in terms of action-angle variables are defined for simply connected domains such as an interior of a homoclinic orbit. On this basis methods of (local) perturbations leading, in particular, to chaotic systems have been studied in literature. We are describing a new method for constructing global action-angle variables and successive perturbations based on a topological covering of the phase space. The method is demonstrated for representative example of the Duffing system.

Motivation & Objective

• To extend the classical action-angle formalism, typically limited to simply connected domains, to topologically nontrivial phase spaces.
• To resolve the limitation of local action-angle variables in Hamiltonian systems with complex topology, such as those with homoclinic orbits.
• To develop a systematic method for constructing global action-angle variables using topological covering and symmetry-based transformations.
• To demonstrate the method’s applicability to the Duffing system, both conservative ($\mu=0$) and dissipative ($\mu\neq0$).

Proposed method

• Utilizes a double-cover transformation mapping the original phase space $(x, y)$ to a new space $(x_1, y_1)$ via $x_1 = x^2 - y^2$, $y_1 = 2xy$, effectively covering the plane twice.
• Employs symmetry under $(x, y) \to (-x, -y)$ to ensure the dynamics are identical on both sheets of the covering space, enabling consistent evolution across the branched structure.
• Introduces a cut along the negative $x$-axis to define distinct sheets (upper and lower planes), with trajectories transitioning smoothly between them upon crossing the cut.
• Rewrites the Duffing system in the new variables $(x_1, y_1)$, yielding a closed-form system (Eq. 4) that respects the covering structure and is independent of the sheet (color) label.
• Transforms the system into polar coordinates $(\rho, \theta)$ in the covering space, with $\theta$ serving as a global angle variable and $\dot{\theta} < 0$ everywhere except at the origin.
• Defines the action variable $I(H)$ globally via the Hamiltonian $H$, ensuring trajectories satisfy standard action-angle equations on the covering space.

Experimental results

Research questions

• RQ1 Can action-angle variables be extended globally to Hamiltonian systems with non-simply connected phase spaces, such as the Duffing system?
• RQ2 How can topological obstructions—such as homoclinic orbits—be overcome to define consistent global action-angle variables?
• RQ3 What role does symmetry play in enabling a global formulation through topological covering?
• RQ4 How does the introduction of dissipation ($\mu \neq 0$) affect the global structure of action-angle variables?
• RQ5 Can the global action-angle formalism support perturbation theory for chaotic systems beyond local approximations?

Key findings

• The transformation $x_1 = x^2 - y^2$, $y_1 = 2xy$ establishes a double covering of the phase space, allowing global definition of action-angle variables by resolving topological obstructions.
• The phase flow in the covering space is independent of the sheet (color), ensuring smooth transition across the cut and enabling a single dynamical system on the branched manifold.
• In the covering space, the angular variable $\theta$ is strictly decreasing ($\dot{\theta} < 0$) everywhere except at the origin, making it a valid global angle variable.
• For $\mu = 0$, trajectories in the covering space are closed curves (circles), satisfying the standard global action-angle equations with $I(H)$ derived from the Hamiltonian.
• For $\mu \neq 0$, $dH/d\theta > 0$ implies that trajectories become spirals, confirming the system's dissipative nature while preserving the global action-angle structure.
• The method generalizes beyond the Duffing system, with a group-theoretic foundation suggesting applicability to a broader class of Hamiltonian systems with symmetry and nontrivial topology.