Hamiltonian & Symplectic Integrable Systems

• Hamiltonian and symplectic integrable systems are defined by maximal sets of commuting first integrals on symplectic manifolds, establishing a framework for Liouville and generalized integrability.
• They exhibit rich local and global invariants such as action-angle variables, monodromy, and integral-affine structures that classify toric and semitoric systems.
• Advanced symplectic integration methods preserve invariant tori and the underlying structure, ensuring long-term stability in numerical simulations of complex dynamical systems.

Hamiltonian and symplectic integrable systems form the cornerstone of modern mathematical physics, providing both the foundation for classical mechanics and deep connections to symplectic geometry, representation theory, and semiclassical spectral theory. These systems are characterized by the presence of maximal sets of commuting first integrals, rich geometric structures, and a robust theory of local and global invariants.

1. Definitions and Structural Framework

A symplectic manifold is a smooth, even-dimensional manifold $(M^{2n},\omega)$ equipped with a closed, nondegenerate 2-form $\omega$. For each smooth function $H\in C^\infty(M)$, the associated Hamiltonian vector field $X_H$ is uniquely determined by

$\iota_{X_H} \omega = dH,$

yielding Hamilton's equations in local Darboux coordinates.

A Hamiltonian system is called completely integrable (in the Liouville sense) if there exist $n$ functions $(f_1,\ldots,f_n)\in C^\infty(M)$ that are:

• Functionally independent almost everywhere: $df_1 \wedge \cdots \wedge df_n \neq 0$ on a dense open set.
• Pairwise in involution with respect to the Poisson bracket induced by $\omega$: $\{f_i,f_j\}=0$ for all $\omega$0.

These functions define a momentum map $\omega$1, whose regular level sets are Lagrangian tori, giving rise to the Liouville foliation of $\omega$2 (Pelayo et al., 2013, Pelayo et al., 2013).

2. Local Normal Forms and Singularities

The local geometry of integrable systems splits into regular and singular cases:

• Regular points: By the Darboux–Carathéodory theorem, there exist local canonical coordinates such that $\omega$3 depend only on half the coordinates (the "actions"), and the symplectic form is standard (Pelayo et al., 2013). This underpins the local existence of action–angle variables $\omega$4, where the flows of all $\omega$5 are linear on the torus and solutions are quasi-periodic:

$\omega$6

• Nondegenerate singularities: Eliasson's normal form theorem classifies isolated nondegenerate critical points via the Williamson types (elliptic, hyperbolic, and focus–focus blocks). In suitable symplectic coordinates, $\omega$7 depends only on canonical quadratic functions corresponding to these block types (Pelayo et al., 2013, Pelayo et al., 2013).

Semiglobal invariants at focus–focus singularities include monodromy and the "Taylor series invariant," reflecting the local symplectic structure of the critical fiber and the nontrivial affine monodromy in the base (Pelayo et al., 2013).

3. Global Invariants and Classification of Integrable Systems

The global classification of integrable systems relies on the affine and convexity structures induced by the momentum map:

• Integral-affine structures: Regular values of $\omega$8 parametrize a base $\omega$9 endowed with an integral-affine atlas (transition maps in $H\in C^\infty(M)$0), canonically arising from action coordinates. Singularities (especially focus–focus) create monodromy—nontrivial affine holonomy around critical values (Pelayo et al., 2013).
• Toric systems: When the Hamiltonian $H\in C^\infty(M)$1-action is effective, the image of the momentum map is a convex polytope (Atiyah–Guillemin–Sternberg theorem). The Delzant theorem asserts that the polytope uniquely determines the triplet $H\in C^\infty(M)$2 up to equivariant symplectomorphism (Pelayo et al., 2013).
• Semitoric systems: On symplectic 4-manifolds, semitoric systems $H\in C^\infty(M)$3—where $H\in C^\infty(M)$4 generates a periodic $H\in C^\infty(M)$5-action and all singularities are nondegenerate and non-hyperbolic—are classified by five symplectic invariants: the number of focus–focus points, Taylor-series invariants, a polygon with cuts generalizing the Delzant polytope, height (volume) invariants, and twisting-index invariants. These invariants completely characterize semitoric systems up to equivalence (Pelayo et al., 2013, Pelayo et al., 2013).

4. Generalizations and Extensions

4.1 Beyond Abelian Integrability

The classical Liouville condition can be relaxed: it suffices for the $H\in C^\infty(M)$6 independent first integrals to generate a solvable Lie algebra under the Poisson bracket. Under appropriate regularity and compatibility conditions, a system with $H\in C^\infty(M)$7 independent, solvably-algebraic first integrals can be solved by $H\in C^\infty(M)$8 successive quadratures, via the method of Lie integrability by quadratures. This includes, but is strictly more general than, Liouville integrability, encompassing certain non-Abelian but solvable symmetry structures (Azuaje, 2023).

4.2 Cosymplectic Perspective

Integrability extends to cosymplectic manifolds $H\in C^\infty(M)$9, relevant for time-dependent Hamiltonian dynamics. Here, a Reeb vector field $X_H$0 (satisfying $X_H$1, $X_H$2) plays the role of time evolution. An Arnold–Liouville-type theorem produces local action–angle coordinates on invariant tori of dimension $X_H$3, with both the symplectic and cosymplectic forms in standard Darboux–type expressions (Jovanovic et al., 2022).

5. Discrete-Time and Symplectic Numerical Integration

Symplectic maps serve as discrete-time analogs of Hamiltonian flows, preserving the symplectic form, and play critical roles in both theoretical dynamics and geometric numerical integration. Integrable symplectic maps admit discrete action–angle variables and possess invariants characterizing phase-space rotation (rotation number) (Zolkin et al., 2017).

For systems with non-separable Hamiltonians, semiexplicit symplectic integrators have been developed that exactly preserve all linear and quadratic invariants. Such schemes (e.g., the Jayawardana–Ohsawa integrator) overcome deficiencies of earlier explicit and extended phase-space methods, enabling structure-preserving simulation in highly general settings and confining numerical trajectories to correct invariant tori (Ohsawa, 2022).

In the context of completely integrable symplectic birational maps, Kahan–Hirota–Kimura (KHK) discretizations produce birational, symplectic, and integrable maps from continuous integrable systems with polynomial Hamiltonians. These maps admit explicit rational first integrals (perturbations of the continuous ones) remaining in involution with respect to the invariant perturbed symplectic structure (Petrera et al., 2016, Petrera et al., 2016).

Closed-form modified Hamiltonians exist for integrable symplectic maps arising from reductions of integrable lattice equations (e.g., discrete KdV), allowing a convergent backward error analysis and demonstrating exponentially excellent long-time integration properties (Alsallami et al., 2017).

6. Persistence and Stability of Invariant Tori

For nearly integrable Hamiltonian systems, symplectic integrators and discrete mappings preserve a large measure set of lower-dimensional elliptic invariant tori, with explicit measure-theoretic estimates on the excluded resonant parameter sets. Under Rüssmann nondegeneracy and small twist conditions, KAM-type theorems hold for both the continuous flow and its symplectic discretization, guaranteeing survival of a "Cantor family" of invariant tori (Shang et al., 2024).

Complementing KAM theory, a Nekhoroshev-type theorem applies to analytic, nearly integrable symplectic maps, ensuring exponential stability of the action variables for exponentially long times, with explicit bounds on drift. This demonstrates strong nonlinear stability and almost-conservation of actions for symplectic algorithms, ensuring practical long-term fidelity in integrable and nearly integrable regimes (Ding et al., 2018).

7. Spectral Theory and Quantum Integrable Systems

Semiclassical quantizations of classical integrable systems yield commuting families of self-adjoint operators. The joint spectrum of such quantum integrable systems asymptotically forms a distorted lattice (in the regular region) or exhibits more involved monodromy and eigenvalue clustering behaviors at singularities (especially focus–focus points). The "inverse spectral conjecture" proposes that, generically, the full joint spectrum determines all the classical symplectic invariants of the system—verified in the toric and one-degree-of-freedom cases, and conjectured for semitoric systems (Pelayo et al., 2013, Pelayo et al., 2013).

---

References:

• "Symplectic theory of completely integrable Hamiltonian systems" (Pelayo et al., 2013)
• "First steps in symplectic and spectral theory of integrable systems" (Pelayo et al., 2013)
• "Lie integrability by quadratures for symplectic, cosymplectic, contact and cocontact Hamiltonian systems" (Azuaje, 2023)
• "Integrable systems in cosymplectic geometry" (Jovanovic et al., 2022)
• "Preservation of Quadratic Invariants by Semiexplicit Symplectic Integrators for Non-separable Hamiltonian Systems" (Ohsawa, 2022)
• "A construction of commuting systems of integrable symplectic birational maps" (Petrera et al., 2016, Petrera et al., 2016)
• "The elliptical invariant tori of nearly integrable Hamiltonian system through symplectic algorithms" (Shang et al., 2024)
• "Rotation number of integrable symplectic mappings of the plane" (Zolkin et al., 2017)
• "Closed-form modified Hamiltonians for integrable numerical integration schemes" (Alsallami et al., 2017)
• "Exponential Stability Estimate of Symplectic Integrators for Integrable Hamiltonian Systems" (Ding et al., 2018)