Pith. sign in

REVIEW 2 major objections 5 minor 19 references

Long time evolution of the H\'enon-Heiles system for small energy

T0 review · 2 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper proves an explicit Poincaré-map formula showing that the Hénon–Heiles slow variables rotate with nearly constant amplitude for times of order ε^{-3}.

desk verdict A serious, likely correct paper on long-time Hénon-Heiles asymptotics; the one load-bearing lemma is a sketch and needs a real proof before the main theorem can be fully trusted. read the letter →

arxiv 2411.16071 v1 pith:WKIPJQJK submitted 2024-11-25 math.CA math-phmath.DSmath.MP

Ovidiu Costin , Rodica Costin , Kriti Sehgal This is my paper · ORCID
classification math.CAmath-phmath.DSmath.MP MSC 34E1037J4070H05
keywords Hénon–HeilessystemslowvariablesPoincarémaplong-timeasymptoticsadiabaticinvariantssmall-energydynamicssecularperturbationtheorytwo-scale
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proves that at small energies the Hénon–Heiles system has a pair of slow observables whose motion is an almost uniform rotation over a time of order $\varepsilon^{-3}$. Specifically, the iterated Poincaré map on the section $y=0$ satisfies $u_n+iw_n=\sqrt{u_0^2+w_0^2}\,e^{i\varphi_n}(1+n\varepsilon^3\eta_n)$ with bounded $\eta_n$, for all $n$ up to order $\varepsilon^{-3}$. This extends far beyond the time horizon $t=O(\varepsilon^{-2})$ at which the ordinary perturbation series develops secular terms. The result gives rigorous asymptotic constants of motion that explain the numerically observed slow drift and filling of the phase-space region. If correct, it provides an explicit formula for long-time evolution in a benchmark non-integrable Hamiltonian system.

What carries the argument

The carrying object is the iterated Poincaré map of the slow variables with respect to the section $y=0$, computed by integrating the complexified equations around the contour $C_{v_0}$ that encircles the two branch points $\pm\sqrt{v_0}$ of $\sqrt{v-y^2}$. The one-loop map produces the exact second-order coefficients displayed in (33), and the reduced recurrence $u_{n+1}=u_n-\varepsilon^2 w_n\sqrt{h^2-u_n^2-w_n^2}-\varepsilon^3 R_n$, $w_{n+1}=w_n+\varepsilon^2 u_n\sqrt{h^2-u_n^2-w_n^2}+\varepsilon^3 S_n$ carries the slow dynamics. The recursion is a rotation by angle $\varepsilon^2\sqrt{h^2-T_n}$ plus a small remainder, and the bootstrap in Lemma 4.4 keeps the remainders $R_n,S_n$ bounded uniformly in $n$ as long as $n\varepsilon^3\le K_0$; this uniform remainder control is what extends validity beyond the perturbation-series horizon.

What would settle it

Take $\varepsilon=0.01$, $h=0.1$, and the initial conditions of Figure 4; integrate with a high-accuracy solver and record $u_n,w_n$ at every return to $y=0$. The theorem predicts $\sqrt{T_n}=\sqrt{T_0}(1+O(n\varepsilon^3))$, so the relative deviation after $n=10^5$ loops must be bounded by a fixed constant times $0.1$; if instead it grows linearly in $n\varepsilon^2$, reaching order $10$ at that $n$, the claim is falsified.

Watch

Extended reading notes

Core claim

The central discovery is that, for initial conditions with $y(0)=0$, $\dot y_0\ne0$, and $u_0^2+w_0^2<h^2$, the slow variables $u=h-(y^2+\dot y^2)$ and $w=\dot x\dot y+xy$ do not drift chaotically on short timescales. Instead, after each return to the section $y=0$ they rotate: $u_n+iw_n=\sqrt{T_0}\,e^{i\varphi_n}(1+n\varepsilon^3\eta_n)$, where $T_0=u_0^2+w_0^2$, $|\eta_n|\le M$, and $\varphi_n$ is given by $\varphi_0+\frac{14\pi}{3}\varepsilon^2\sum_{k=0}^{n-1}\sqrt{h^2-T_k}$. For $n\varepsilon^{5/2}\ll1$ the phase simplifies to $\varphi_0+n\varepsilon^2\sqrt{h^2-T_0}+O(n^2\varepsilon^5)$, so the radius $\sqrt{T_n}$ is conserved up to $O(n\varepsilon^3)$ while the angle advances at a slowly varying rate. The proof runs through a Poincaré map in complex time, integrating once around the branch points $y=\pm\sqrt{v_0}$; the one-loop change of the slow variables is exactly of order $\varepsilon^2$ with coefficients $\frac{14\pi}{3}w_0\sqrt{2hv_0-v_0^2-w_0^2}$ and $\frac{14\pi}{3}(h-v_0)\sqrt{2hv_0-v_0^2-w_0^2}$, and the $\varepsilon^3$ remainder is controlled uniformly for $N\varepsilon^3\le K_0$ by a bootstrap.

Load-bearing premise

The load-bearing premise is that the $\varepsilon^3$ remainder terms $R_n,S_n$ in the slow-variable recurrence stay bounded by a constant depending only on the initial distance from the boundary $h^2$, uniformly over all loops with $N\varepsilon^3\le K_0$; the whole theorem rests on that bootstrap succeeding.

Editorial extensions

If this is right

  • The standard $\varepsilon$-expansion of $x(t),y(t)$ develops secular terms when $t\varepsilon^2=O(1)$; the theorem implies the slow-variable rotation remains accurate for $n$ of order $\varepsilon^{-3}$, extending the reliable time horizon by a factor of $\varepsilon^{-1}$.
  • For $n\varepsilon^{5/2}\ll1$, formulas (16) and (58) give explicit trigonometric expressions for $u_n$ and $w_n$, so long-time evolution can be predicted without solving the differential equations over that entire interval.
  • The squared radius $T_n=u_n^2+w_n^2$ is an approximate adiabatic invariant: it changes by at most $O(n\varepsilon^3)$, and in the simplified regime the amplitude of the slow oscillation is constant to that accuracy.
  • In the degenerate cases $\dot y_0=0$ or $u_0=w_0=0$, the slow variables stay $O(n\varepsilon^3)$, so the system exhibits no slow rotation beyond the trivial drift.
  • The numerical comparisons in Section 5 show the formulas tracking $v_n$ and $w_n$ over more than a million loops at $\varepsilon=0.01$, in agreement with the predicted error size.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural next step, not taken in the paper, is to carry the expansion one order higher: if the remainders can be controlled at $\varepsilon^4$, the same contour method should yield validity up to $n\varepsilon^4=O(1)$ with a corrected phase.
  • Because the proof only uses the rational structure of the integrands in $x$, $\sqrt{v-y^2}$, and $S$, the same two-scale Poincaré-map construction may apply to other two-degree-of-freedom Hamiltonians with cubic nonlinearities and a resonant harmonic limit.
  • The conservation of $T_n$ suggests that the level sets $T=\text{constant}$ act as approximate invariant tori for small $\varepsilon$; a direct numerical test would be to measure how long a trajectory stays within $O(\varepsilon)$ of such a level set.
  • The phase formula (14) depends only on $T_k$, so it can be iterated as a cheap map even when the simplified formula (16) is no longer accurate; comparing the two against numerics would locate the exact $n\varepsilon^{5/2}$ threshold.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper studies the Hénon–Heiles system in the small-energy regime after rescaling, and derives long-time asymptotic formulas for the slow variables u = h - (y^2 + \dot y^2) and w = \dot x\dot y + xy. The main result, Theorem 3.1(iii), states that for initial data with u_0^2 + w_0^2 < h^2, the iterated Poincaré map of the slow variables satisfies u_n + i w_n = \sqrt{u_0^2+w_0^2} e^{i\varphi_n}(1 + n\varepsilon^3\eta_n) with an explicit phase \varphi_n and uniformly bounded error, for n up to \varepsilon_0^{-3}. For n\varepsilon^{5/2}\ll 1 the phase reduces to a uniform rotation. The proof uses a complex-plane Poincaré map, an \varepsilon-expansion with remainders, and a bootstrap argument; the paper also includes explicit second-order expansions and a numerical comparison.

Significance. If the proof gap discussed below is filled, the result is significant: it gives rigorous asymptotic control of the slow dynamics on a timescale n\sim\varepsilon^{-3}, which goes well beyond the standard perturbation horizon n\varepsilon^2=O(1). The derivation is self-contained and has no fitted parameters; the explicit formulas in Appendices B and C and the independent numerical check in Figures 4 and 6 are genuine strengths. The main bottleneck is the proof of Lemma 4.2, which is sketched rather than proved and is load-bearing for Theorem 3.1(iii).

major comments (2)
  1. [4.5, Lemma 4.2] The proof of Lemma 4.2 is only a two-sentence sketch ('the structure of the integral operators ... is similar', 'the same arguments go through'), but this lemma is the sole source of the uniform remainder bounds |R_n|, |S_n|\le M used in the bootstrap Lemma 4.4. A complete proof needs to establish several quantitative facts that are not immediate from Lemma 4.1: (i) uniform sup-norm bounds on the source terms f_1,g_1 in (31), which contain v[2], w[2] from (84)-(85) with arcsin(y/\sqrt{v_0}) and (v_0-y^2)^{\pm 3/2} along the two-circle contour C_{v_0}; (ii) a contraction estimate for (I_1,I_2) in an O(1)-ball of the Banach space B, noting that R,S are O(1), not small in \varepsilon, so the Taylor remainder Q_F,Q_G in (30) must be controlled throughout that ball; (iii) uniformity of all constants in the initial data (v_0,w_0) ranging over the annulus c_1\le (h-v_0)^2+w_0^2\le c_0^2, since Lemma 4.4 re-applies Lemma 4.2 about \varepsilon_0^{-3} times. Without this uniformity the bootstrap in Lemma 4.4, and with it Theorem 3.1(iii), does not follow as written.
  2. [3.2, Theorem 3.1(i)] The statement 'if \dot y_0=0, then w_n=O(n\varepsilon^3) and u_n=O(n\varepsilon^4)' is inconsistent with the definition u=h-v. Since \dot y_0=0 gives v_0=0 and hence u_0=h, the claim u_n=O(n\varepsilon^4) fails already at n=0 for every h>0. Appendix D likewise gives v(t)=O(\varepsilon^2)+O(t\varepsilon^4), so the correct statement is of the form u_n=h+O(\varepsilon^2)+O(n\varepsilon^4), not u_n=O(n\varepsilon^4). Please correct the theorem statement and the corresponding proof summary in \S4.7.1.
minor comments (5)
  1. [4.4-4.5] The paragraphs 'Solutions along one loop' and 'Solutions Along One Loop' are duplicated nearly verbatim; one copy should be removed.
  2. [4.7.3-4.7.4] Both subsections are titled 'Proof of (iii)'; the second should refer to the simplified phase formula (16) or to the n\varepsilon^{5/2} regime, not repeat the same title.
  3. [4.6, around Eq. (69)] The displayed recursion for B_n is typeset in a way that is hard to parse: the product index, the \sqrt{T_0} factor, and the relation between \tilde C_\ell and C_\ell should be written out explicitly, because the telescoping identity \prod \tilde C_\ell=(\sqrt{T_0}/\sqrt{T_n})\prod C_\ell is used implicitly to pass to (71).
  4. [5, Figure 4] The caption reports n=1024000 and \varepsilon=0.01, giving n\varepsilon^3=1.024; since the theorem guarantees the estimates for N\varepsilon_0^3\le K_0, please state the value of K_0 for these parameters or note explicitly that the plotted data lie in the asymptotic (not necessarily proven) regime.
  5. [3.2, Eq. (14)] In the sentence immediately after (16), 'Formula (14) simplifies to (16)' is correct, but the line 'For n slightly smaller, such that n\varepsilon^{5/2}\ll 1' should be reconciled with the qualitative claim in Remark 3.2 that this range can still be much larger than the secular range n\varepsilon^2=O(1); a sentence clarifying the ordering would help.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the iterated Poincaré-map recurrence (49) and phase formula (14) are computed from the equations of motion with explicit, unfitted coefficients; the only self-citations ([10],[11]) attribute the general method and are not load-bearing.

full rationale

The derivation chain is self-contained. The slow variables u = h − (y² + ẏ²) and w = ẋẏ + xy are defined directly from the rescaled equations (4); after the change of variables to (h, v, w, t) with independent variable y (Eqs. (18)–(21)), the one-loop Poincaré map (33) is obtained by the explicit ε-expansion (28), with v[1], w[1], v[2], w[2] computed in closed form (Appendices B–C). The coefficient 14π/3 arises from the arcsin monodromy gain of 2π around the two-branch-point contour C_{v0} (§4.2.2), not from any fitted parameter. Iterating (33) yields the recurrence (49), and Lemma 4.3 derives the amplitude/phase formulas (13)–(16) from (49) under the uniform bounds |R_n|, |S_n| ≤ M; Lemma 4.4 obtains those bounds by a bootstrap that re-applies Lemma 4.2 at each cycle while T_n stays in the annulus c₁ ≤ T_n ≤ c₀². No parameter is fitted: ε, h, and the initial conditions are the physical inputs, and ε₀, K₀, M are existence constants, never matched to data. The numerical comparison in §5 is an independent check of formula (17) against direct ODE integration, with error reported in Figure 6. The only self-citations are methodological: §2.3 ('we utilize "approximate adiabatic invariants", an approach and methods introduced in [10], [11]') and §3.1 ('As in [10], [11], we use the Poincaré map to eliminate the fast variable'); no lemma, bound, or uniqueness claim is imported from those papers, so they are not load-bearing. The most delicate step, Lemma 4.2, states the O(ε³) remainder bounds with a two-sentence proof ('The same arguments as in the proof of Lemma 4.1 go through straightforwardly'), which is a completeness and rigor concern about uniformity across loops, not a circularity: the bound is claimed from a contraction argument on the system (31), not assumed from the theorem being proved.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The main formulas contain no fitted free parameters: ε, h, and initial conditions are inputs, and the constants ε0, K0, M in Theorem 3.1 are existence bounds, not chosen to match data. The proof rests on standard analytic ODE tools and two domain assumptions: that the complex contour Cv0 integration represents one real oscillation, and that the initial slow action lies strictly inside the allowed disk (10). No new physical entities are introduced.

assumptions (4)
  • domain assumption Analyticity of solutions of (4) in t, ε and initial conditions, and the ability to continue them along the complex loop Cv0.
    Used in Section 4.1 and Lemma 4.1; the Hénon-Heiles vector field is polynomial, so real analytic solutions extend locally, but the global continuation along the chosen contour is asserted.
  • standard math Contraction mapping theorem and standard analytic estimates.
    Used in Lemma 4.1 to prove existence and uniqueness of the solution on one loop.
  • domain assumption The loop integral around Cv0 generates the Poincaré map of the real system on y=0.
    Section 4.1, equations (18)-(27) and Corollary 4.0.1; this identification is standard for pendulum-type equations but is a non-trivial modeling step.
  • domain assumption Initial condition constraint (10): u0^2 + w0^2 < h^2.
    Theorem 3.1(iii) excludes the boundary, ensuring denominators stay bounded; used in Lemmas 4.1 and 4.4.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Long time evolution of the H\'enon-Heiles system for small energy." pith.science (2026). https://pith.science/paper/WKIPJQJK

@misc{pith2026241116071,
  author       = {Pith},
  title        = {Pith review of: Long time evolution of the H\'enon-Heiles system for small energy},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WKIPJQJK}},
  note         = {Machine review of arXiv:2411.16071}
}
read the original abstract

The H\'enon-Heiles system, initially introduced as a simplified model of galactic dynamics, has become a paradigmatic example in the study of nonlinear systems. Despite its simplicity, it exhibits remarkably rich dynamical behavior, including the interplay between regular and chaotic orbital dynamics, resonances, and stochastic regions in phase space, which have inspired extensive research in nonlinear dynamics. In this work, we investigate the system's solutions at small energy levels, deriving asymptotic constants of motion that remain valid over remarkably long timescales -- far exceeding the range of validity of conventional perturbation techniques. Our approach leverages the system's inherent two-scale dynamics, employing a novel analytical framework to uncover these long-lived invariants. The derived formulas exhibit excellent agreement with numerical simulations, providing a deeper understanding of the system's long-term behavior.

Figures

Figures reproduced from arXiv: 2411.16071 by the authors.

Figure 1
Figure 1. Numerically calculated evolution of the system for initial conditions x(0) = 0.1, y(0) = 0, x˙(0) = 0.08, and y˙(0) = 0.1. d 2x dt2 = −x − 2εxy, d 2y dt2 = −y + εy2 − εx2 . (4) The Hamiltonian rescales as h 7→ h/ε2 ; for simplicity, we will continue to denote it by h: (5) h = 1 2 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Comparison of the evolution of the system with initial conditions x(0) = 0.1, y(0) = 0, x˙(0) = 0.08, and y˙(0) = 0.1 from time t = 0 to t = 1000 for various values of ε. 3 [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The evolution of the slow variables v(t) and w(t), and a non-slow variable x(t) with initial conditions x0 = √ 3/5 2 , y0 = 0, x˙ 0 = 1 5 , y˙0 = 1 10 and ε = 0.1. 3.2. Main results. The main theoretical result is Theorem 3.1. It demonstrates that the iterated Poincaré map of the slow variables u, w with respect to the manifold y = 0 satisfies the recursive formula (13), (14) (in real variables (17)), which is valid… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Comparison between numerical calculations and (17) for n = 1024000, h = 0.1, ε = 0.01, x0 = √ 3/5 2 , y0 = 0, x˙ 0 = 0.2, and y˙0 = 0.1. 1The plot in [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: The curve of integration Cv0 To be precise, we consider the path of integration to be the curve Cv0 defined as follows: starting at y = 0 the path goes counterclockwise along a circle centered at √ v0 and of radius √ v0 followed, counterclockwise, by the circle centere…
Figure 6
Figure 6. Figure 6: Error Analysis: maximum value of the difference between the numerical and theoretical value of vn on the interval [0, n] as a function of n. Appendix A. Solution through perturbation expansion The ε term in the perturbation expansion for x(t) and y(t) are xε(t) = − 4 3…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references · 19 canonical work pages

  1. [2]

    and Serrano, S., 2008

    Barrio, R., Blesa, F. and Serrano, S., 2008. Fractal structures in the Hénon-Heiles hamiltonian. Europhysics Letters, 82 (1), p.10003

    work page 2008

  2. [1]

    Geometrical methods in the theory of ordinary differential equations (Vol

    Arnold, V.I., 2012. Geometrical methods in the theory of ordinary differential equations (Vol. 250). Springer Science & Busi- ness Media

    work page 2012

  3. [3]

    Linearization of resonant vector fields

    Basto-Gonçalves, J., 2010. Linearization of resonant vector fields. Transactions of the American Mathematical Society, 362(12), pp.6457-6476

    work page 2010

  4. [4]

    and Sanjuan, M.A., 2012

    Blesa, F., Seoane, J.M., Barrio, R. and Sanjuan, M.A., 2012. To escape or not to escape, that is the question—perturbing the Hénon–Heiles Hamiltonian. International Journal of Bifurcation and Chaos, 22 (06), p.1230010

    work page 2012

  5. [5]

    Stochastic behavior in classical and quantum hamiltonian systems

    Casati, G., 1979. Stochastic behavior in classical and quantum hamiltonian systems. Lecture notes in physics

    work page 1979

  6. [6]

    and Weiss, J., 1982

    Chang, Y.F., Tabor, M. and Weiss, J., 1982. Analytic structure of the Hénon–Heiles Hamiltonian in integrable and noninte- grable regimes. Journal of Mathematical Physics, 23 (4), pp.531-538

    work page 1982

  7. [7]

    and Rod, D.L., 2005, June

    Churchill, R.C., Pecelli, G. and Rod, D.L., 2005, June. A survey of the Hénon-Heiles Hamiltonian with applications to related examples. In Stochastic Behavior in Classical and Quantum Hamiltonian Systems: Volta Memorial Conference, Como, 1977 (pp. 76-136). Berlin, Heidelberg: Springer Berlin Heidelberg

    work page 2005

  8. [8]

    and Teichmann, T., 1956

    Coddington, E.A., Levinson, N. and Teichmann, T., 1956. Theory of ordinary differential equations

    work page 1956

Show all 19 references
  1. [9]

    and Verhoeven, C., 2005

    Conte, R., Musette, M. and Verhoeven, C., 2005. Explicit integration of the Hénon-Heiles Hamiltonians. Journal of Nonlinear Mathematical Physics, 12 (sup1), pp.212-227

    2005

  2. [10]

    and Huang, M., 2016

    Costin, O., Costin, R.D. and Huang, M., 2016. A direct method to find Stokes multipliers in closed form for P1 and more general integrable systems. Transactions of the American Mathematical Society, 368 (11), pp.7579-7621

    2016

  3. [11]

    and Huang, M., 2015

    Costin, O., Costin, R.D. and Huang, M., 2015. Tronquée solutions of the Painlevé equation PI. Constructive Approximation, 41, pp.467-494. 21

    2015

  4. [12]

    and Letelier, P.S., 1999

    de Moura, A.P. and Letelier, P.S., 1999. Fractal basins in Hénon–Heiles and other polynomial potentials. Physics Letters A, 256(5-6), pp.362-368

    1999

  5. [13]

    The Hénon-Heiles system revisited

    Fordy, A.P., 1991. The Hénon-Heiles system revisited. Physica D: Nonlinear Phenomena, 52 (2-3), pp.204-210

    1991

  6. [14]

    and Heiles, C., 1964

    Hénon, M. and Heiles, C., 1964. The applicability of the third integral of motion: some numerical experiments. Astronomical Journal, Vol. 69, p. 73 (1964), 69, p.73

    1964

  7. [15]

    Non-integrability of Hénon-Heiles system and a theorem of Ziglin.Kodai mathematical journal, 8(1), pp.120-138

    Ito, H., 1985. Non-integrability of Hénon-Heiles system and a theorem of Ziglin.Kodai mathematical journal, 8(1), pp.120-138

    1985

  8. [16]

    and Stuart, A., 2008

    Pavliotis, G.A. and Stuart, A., 2008. Multiscale methods: averaging and homogenization (Vol. 53). Springer Science & Business Media

    2008

  9. [17]

    and Caboz, R., 1993

    Ravoson, V., Gavrilov, L. and Caboz, R., 1993. Separability and Lax pairs for Hénon–Heiles system. Journal of mathematical physics, 34(6), pp.2385-2393

    1993

  10. [18]

    and Du, M.L., 2007

    Zhao, H.J. and Du, M.L., 2007. Threshold law for escaping from the Henon-Heiles system. Physical Review E—Statistical, Nonlinear, and Soft Matter Physics, 76 (2), p.027201

    2007

  11. [19]

    Classifying orbits in the classical Hénon–Heiles Hamiltonian system

    Zotos, E.E., 2015. Classifying orbits in the classical Hénon–Heiles Hamiltonian system. Nonlinear Dynamics, 79 , pp.1665- 1677. 22

    2015

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.