Study of Duffing oscillator using an improved Lindstedt Poincare method and relevant comparisons

Ramij Ahamed; Subhankar Ray

REVIEW 3 major objections 5 minor 12 references

A modified Lindstedt–Poincaré method that precomputes the exact frequency of the undamped Duffing oscillator converges to the numerical solution after only fourth order, matching the accuracy that standard LPM reaches at tenth order and Bur

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 23:46 UTC pith:LHHUNYQH

load-bearing objection LPM-M as written is standard LPM plus a final exact-frequency substitution; the elliptic-integral frequency derivation is correct but the convergence claim is not well-defined and needs clarification and error metrics. the 3 major comments →

arxiv 2607.15233 v1 pith:LHHUNYQH submitted 2026-07-16 math-ph math.MPphysics.class-phphysics.comp-ph

Study of Duffing oscillator using an improved Lindstedt Poincare method and relevant comparisons

Ramij Ahamed , Subhankar Ray This is my paper

classification math-ph math.MPphysics.class-phphysics.comp-ph MSC 34E1034C15

keywords Duffing oscillatorLindstedt-Poincaré methodPerturbation theoryExact frequencyJacobi elliptic functionSecular termsNonlinear oscillatorConvergence

verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper proposes an improved Lindstedt–Poincaré method (LPM-M) for the undamped, unforced Duffing oscillator. Its central claim is that inserting the exact frequency—obtained from an elliptic-integral formula—into the perturbation hierarchy removes secular terms from the first order onward, so solutions stay in phase with numerical integration over many periods. The authors show that fourth-order LPM-M matches the accuracy of tenth-order standard LPM and eighth-order Burton LPM at a representative parameter set, and that LPM-M exhibits no phase lead or lag at large times. A sympathetic reader would care because perturbation methods are the workhorse for nonlinear oscillator analysis, and a method that needs far fewer orders and eliminates phase drift is practically valuable.

Core claim

The paper's central discovery is that the slow convergence and phase drift of the standard Lindstedt–Poincaré method come largely from using truncated frequency corrections. If one first computes the exact angular frequency ω_ex = (π/2)√(1+εA²)/K(εA²/(2(1+εA²))) using the complete elliptic integral K, expands it in powers of ε, and then uses that known series (with coefficients ν_n that happen to equal the standard LPM corrections ω_n) in the perturbation equations, the same solution hierarchy results but with no resonance terms at first order. Evaluating the final perturbation sum at ω_ex t instead of at the truncated frequency keeps the solution in phase with the exact motion; the paper de

What carries the argument

The key machinery is the exact frequency derived from the first integral of the Duffing equation: ω_ex = (π/2)√(1+p)/K(p/(2(1+p))) with p=εA². This exact frequency is expanded as a power series ω_ex = Σ εⁿνₙ, and the perturbation problem is reformulated with the stretched time τ₂=ω_ex t while keeping the truncated series ν[N] in the differential equation to cancel secular terms. The mechanism that carries the argument is the evaluation of the perturbation solution at the exact frequency rather than at an order-by-order corrected frequency, which is what preserves the phase at all times.

Load-bearing premise

The method's superiority over standard LPM depends on knowing the exact frequency ω_ex before solving, which exists here because the Duffing equation has a closed-form energy integral; for a general nonlinear oscillator no such exact frequency is available, so the claimed versatility hinges on that a priori knowledge.

What would settle it

A concrete falsifier: take a nonlinear oscillator with no closed-form exact frequency (e.g., x'' + x + εx³ + δx⁵), attempt LPM-M by approximating ω_ex numerically, and check whether the fourth-order phase preservation still holds; alternatively, recompute the paper's LPM-M solution by evaluating at ν[N] t (as §4.2 describes) instead of ω_ex t and show that the phase error reappears, which would demonstrate that the improvement comes solely from the final frequency substitution.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

If the claim is correct, LPM-M provides accurate long-time Duffing solutions with only fourth-order expansions, reducing the algebraic complexity of perturbation calculations dramatically.
The method eliminates phase drift entirely when the exact frequency is known, which is the dominant error source in standard LPM at large times.
The equality of the frequency-series coefficients (ν_n = ω_n) implies that the improvement is a re-summation effect: the perturbation solution coefficients are the same, but the frequency resummation changes the evaluation time base.
For engineering applications, the method offers a practical way to obtain reliable oscillator waveforms from a handful of terms rather than the 8–10 terms required by earlier methods.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The method's generality is likely limited to oscillators whose exact frequency (or period) can be written in closed form, since the entire advantage rests on precomputing ω_ex; for systems without such a formula the method reduces to standard LPM.
Because the solution coefficients y_n are identical to the standard LPM x_n, the improvement could be replicated by a simple Padé-like resummation of the standard frequency series, which is a testable hypothesis the paper does not explore.
A quantitative error norm (e.g., L2 difference from numerical solution over many periods) would sharpen the visual convergence claim, and one would expect the phase error to be the dominant term suppressed by the exact frequency.
The approach may extend naturally to other integrable or near-integrable oscillators (e.g., the pendulum) where exact periods are known via elliptic integrals, providing a concrete next test.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

LPM-M as written is standard LPM plus a final exact-frequency substitution; the elliptic-integral frequency derivation is correct but the convergence claim is not well-defined and needs clarification and error metrics.

read the letter

Short take: this paper has a real but small idea—use the exact Duffing frequency from the elliptic integral in the final argument of a standard Lindstedt–Poincaré solution—but as written the method is not well-defined and the central convergence claim is not supported.

What is genuinely good: the exact frequency derivation from the energy integral (eq. 38) is clean, the perturbation coefficients through 10th order are correct, and Figures 2–5 do show the phase-error problem clearly. The authors are right that LPM and LPM-B frequencies oscillate around the exact value and that phase error grows with time.

The soft spots are significant. Section 4.2 sets up the perturbation equations with the truncated frequency ν[N]—which, as eq. (41) shows, is exactly the standard LPM frequency expansion—and then eq. (60) evaluates the solution at ω_ex t. Since ν_i = ω_i, solving with ν[N] and evaluating at ν[N] t would reproduce standard LPM identically. The entire observed improvement comes from that final substitution of ω_ex, and the paper never justifies it or explains why the perturbation expansion derived with ν[N] can be evaluated at a different time scale. The conclusion even claims ω_ex is 'used in all orders of calculation,' which contradicts §4.2. That internal inconsistency is not a minor typo; it defines what the method actually is.

The evidence is also only visual, for a single parameter pair (ε=0.4, A=1.5), with no quantitative error norms. And the comparison is not fair: LPM-M is handed the exact frequency as an external input, while LPM and LPM-B have to estimate it from the series. Of course the phase error disappears. That doesn't demonstrate better convergence in a nontrivial sense.

The 'highly versatile' claim is overstated. For most nonlinear oscillators an exact elliptic-integral frequency is not available, so the trick does not generalize as advertised.

Who gets value from this? Someone working on analytic approximations for Duffing-type oscillators might find the phase-correction idea useful for a specific application, but the paper needs a clear statement of the method (which frequency goes into the ODE and which goes into the final time argument), a quantitative error comparison across ε and A, and a more modest framing.

My recommendation: send it to peer review only if the authors are willing to make those revisions. The derivation work is real, and a good referee could push it into a useful short paper. As it stands, the main claim is not reproducible.

Referee Report

3 major / 5 minor

Summary. The paper studies the undamped, unforced Duffing oscillator (Eq. 1) and compares three perturbation approaches: the standard Lindstedt–Poincaré method (LPM), Burton's modification (LPM-B, expanding ω²), and a proposed 'Modified LPM' (LPM-M) that uses the exact frequency obtained from the elliptic-integral time period. The authors derive the exact frequency (Eq. 38), expand it in ε, and then use this frequency to construct the perturbation solution. They claim that LPM-M converges much faster than LPM and LPM-B, based on visual agreement with high-precision numerical solutions at ε=0.4, A=1.5.

Significance. If the proposed LPM-M is well-defined and its claimed convergence advantage is quantitatively demonstrated, the paper would offer a useful comparison of perturbation methods for the Duffing oscillator and a practical recipe for improving accuracy when an exact frequency is known. The exact-frequency derivation via the elliptic integral (Section 4.1) is a genuine strength, and the coefficients of the expansion are consistent with standard LPM at each order. However, the central methodological novelty and the headline convergence claim currently rest on an ambiguous procedural description and on visual evidence from a single parameter set. The paper's broader claim that LPM-M is 'highly versatile' for a wide variety of nonlinear oscillators is not supported, as the method depends on knowing the exact frequency a priori, which is not generally available.

major comments (3)

[§5, Figures 2–5] The claim that LPM-M 'converges better' than LPM and LPM-B is not supported by any quantitative error measure. The evidence is purely visual, and it is presented for a single parameter set (ε=0.4, A=1.5) at specific time intervals. The paper should provide a defined error norm (e.g., maximum absolute error or RMS error over a fixed interval) for each method and order, and ideally for multiple values of ε and A, to substantiate the headline conclusion. Without such data, the claim of superior convergence remains anecdotal.
[§6, Conclusions] The concluding claim that LPM-M is 'highly versatile' and offers an effective analytical tool for 'a wide variety of nonlinear oscillators' is not supported by the analysis. The proposed method relies on knowing the exact frequency ω_ex a priori, which for the Duffing oscillator is derived from the elliptic integral in Eq. (38). For a general nonlinear oscillator, such an exact closed-form frequency is not generally available, and the paper provides no procedure for constructing it. The versatility claim should be removed or severely qualified, or a general method for obtaining ω_ex should be given.
[§4.2, Eq. (49)] There is a likely typographical error in the O(ε^5) equation: the term '+6y0y1y3' should presumably be '−6y0y1y3' to match the corresponding standard LPM equation (9), and the term '−2ν2ν3y0' should be '−2ν2ν3¨y0' (double derivative). While this does not affect the reported lower-order results, it obscures the method's implementation and should be corrected.

minor comments (5)

[§4.2, Eq. (60)] The text says 'From (60)' but the displayed equation is Eq. (60) itself; it should refer to Eq. (43).
[§2, Eq. (10)] The O(ε^10) equation is only partially displayed with ellipses; if it is not intended to be fully written, a statement of the general structure would be clearer. Also, in Eq. (24) the amplitude is written with lowercase 'a' (a^20) inconsistently with the rest of the paper.
[§3, Eq. (33)] The expression for α8 and z8 is truncated; please provide the full term or a reference to a supplementary file if exact expressions are needed.
[References] Reference [4] for Lindstedt is incomplete (missing full title/pages). Reference [2] also lacks publisher details; please complete the bibliographic entries.
[§5, Figures] The figures would benefit from clear legends inside each panel and from consistent axis labeling. In Figure 5, the order labels in the text ('10th order LPM, 8th order LPM-B, 4th order LPM-M') should be explicitly matched to the line styles in the figure.

Circularity Check

1 steps flagged

LPM-M's superior convergence is built in: the perturbation part reproduces standard LPM (ν_i=ω_i), and the only difference is evaluating at the externally supplied exact frequency ω_ex.

specific steps

other [§4, eqs. (41), (42), (60); §5 comparison]
"We see that ν0 = 1 , ν1 = ω1, ν2 = ω2, ν3 = ω3, ν4 = ω4, ν5 = ω5 and so on. ... Here, we put the truncated frequency (ν[N]) in place of exact frequency (ωex) to remove the secular terms. ... From (60) y[5](τ2) = Σ5 i=0 ϵ^i y_i(ωext)"

Because ν_i = ω_i, the O(ε^i) equations of LPM-M are identical to those of standard LPM, so y_i(τ) = x_i(τ). The only difference in the plotted solution is the replacement of the truncated LPM frequency by ω_ex in eq. (60). Hence the claimed 'remarkably good convergence by fourth order' is not derived by the perturbation procedure; it is the externally supplied exact frequency inserted by construction. If ν[N] were kept throughout, the method would be exactly standard LPM. The text is also inconsistent: it says ν[N] is used to remove secular terms, but then evaluates at ω_ex t, so the improvement rests on an unexplained final substitution.

full rationale

There are no self-citations, and the exact frequency ω_ex is independently obtained from the energy integral / elliptic integral (eqs. 34–38), so the frequency expansion coefficients themselves are not circularly determined. However, the paper's central claim—that LPM-M converges better than LPM/LPM-B—reduces by construction to inserting that exact frequency into the standard LPM series: since ν_i = ω_i, the LPM-M perturbation equations solve to the same x_i, and only the time argument is changed to ω_ex t. The claimed convergence advantage is therefore an input, not a prediction of the modified method. This is a partial circularity (score 6), not a full one, because the amplitude coefficients are still independently computed and the exact frequency is not fitted to the numerical solution used for comparison. The additional claim that the method is 'highly versatile' for general nonlinear oscillators is also unsupported, but that is a limitation rather than circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 3 axioms · 0 invented entities

The paper introduces no new entities or fitted parameters. It relies on the exact frequency from the elliptic integral (an external result) and assumes the asymptotic validity of the exact-frequency substitution. No numerical fitting is involved.

axioms (3)

domain assumption The exact angular frequency ω_ex = (π/2)√(1+p)/K(p/(2(1+p))) is the true frequency of the Duffing oscillator.
Derived from the energy integral and standard elliptic integral identities, but imported into the perturbation method as a known input; the method's superiority relies on this external knowledge.
standard math The perturbation coefficients ν_i, from expanding ω_ex in powers of p=εA², are identical to the LPM frequency corrections ω_i for all i.
Shown in §4.2 by direct Taylor expansion; this is an identity, not an empirical assumption, but it is load-bearing for the claim that LPM-M's equations match LPM's.
domain assumption The truncated series y[N](ω_ex t) = Σ ε^i y_i(ω_ex t) is a valid asymptotic approximation of order ε^N to the true solution.
The paper does not prove this; it assumes that substituting the exact frequency preserves the asymptotic order of the perturbation solution. No error estimate is provided.

pith-pipeline@v1.3.0-alltime-deepseek · 7926 in / 13949 out tokens · 105724 ms · 2026-08-01T23:46:43.748702+00:00 · methodology

0 comments

read the original abstract

The undamped Duffing oscillator is a nonlinear dynamical system with broad applications in physics, engineering and biological system. We present a comprehensive analysis of this system using the Lindstedt Poincare method (LPM) and its modifications and make comparison with numerical solution obtained using higher order Runge-Kutta. It is also shown the method suggested in this article converges better than the standard LPM and Lindstedt Poincare method with Burton's modification.

Figures

Figures reproduced from arXiv: 2607.15233 by Ramij Ahamed, Subhankar Ray.

**Figure 1.** Figure 1: Frequency for different orders of LPM and LPM-B numerical LPM (4th odr) LPM-B (4th odr) LPM-M (4th odr) T=4.8714 8T 8.2T 8.4T 8.6T 8.8T 9T -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 t x(t) [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗

**Figure 2.** Figure 2: x[4](t) comparison in lower time (8T ≤ t ≤ 9T) [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

**Figure 5.** Figure 5: x(t) in 10th order with LPM, 8th order with LPM-B and 4th order with LPMM for (20T ≤ t ≤ 20.5T) numerical LPM (4th odr) LPM-B (4th odr) LPM-M (4th odr) T=4.8714 8T 8.2T 8.4T 8.6T 8.8T 9T -2 -1 0 1 2 t x'(t) [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗

**Figure 4.** Figure 4: Comparison of x[5](t) at higher time (20T ≤ t ≤ 21T) Moreover, the standard LPM and LPM-B solutions lag in phase for odd orders(Figures 2 and 3) while for even orders the solutions lead in phase as is shown in [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗

discussion (0)

Reference graph

Works this paper leans on

12 extracted references

[1]

Nonlinear dynam- ics and chaos: with applications to physics, biology, chemistry, and engi- neering (studies in nonlinearity) , vol- ume 1

Steven H Strogatz. Nonlinear dynam- ics and chaos: with applications to physics, biology, chemistry, and engi- neering (studies in nonlinearity) , vol- ume 1. Westview press, 2001

2001
[2]

Erzwungene Schwingun- gen bei ver¨ anderlicher Eigenfrequenz und ihre technische Bedeutung

Georg Duﬃng. Erzwungene Schwingun- gen bei ver¨ anderlicher Eigenfrequenz und ihre technische Bedeutung . Number 41-
[3]

Non- linear oscillations

Ali H Nayfeh and Dean T Mook. Non- linear oscillations . John Wiley & Sons, 2024

2024
[4]

Lindstedt

A. Lindstedt. Mem. Acad. Imper. Sci. St. Petersburg, 31, 1883
[5]

Ho- motopy analysis approach to duﬃng- harmonic oscillator

Shao-dong Feng and Li-qun Chen. Ho- motopy analysis approach to duﬃng- harmonic oscillator. Applied Mathe- matics and Mechanics , 30(9):1083–1089, 2009

2009
[6]

Applica- tion of he’s energy balance method to duﬃng-harmonic oscillators

M Momeni, N Jamshidi, Amin Barari, and Davood Domiri Ganji. Applica- tion of he’s energy balance method to duﬃng-harmonic oscillators. Interna- tional Journal of Computer Mathemat- ics, 88(1):135–144, 2011

2011
[7]

He’s parameter-expanding methods for strongly nonlinear oscilla- tors

Lan Xu. He’s parameter-expanding methods for strongly nonlinear oscilla- tors. Journal of Computational and Applied Mathematics , 207(1):148–154, 2007

2007
[8]

Perturbation methods

Ali H Nayfeh. Perturbation methods . John Wiley & Sons, 2024

2024
[9]

Non- linear ordinary diﬀerential equations: an introduction for scientists and engineers

Dominic Jordan and Peter Smith. Non- linear ordinary diﬀerential equations: an introduction for scientists and engineers . Oxford University Press, 2007

2007
[10]

A perturbation method for certain non-linear oscillators

TD Burton. A perturbation method for certain non-linear oscillators. Interna- tional Journal of Non-Linear Mechanics , 19(5):397–407, 1984

1984
[11]

Strong nonlinear oscillators

Livija Cveticanin. Strong nonlinear oscillators. Mathematical Engineering. Analytical Solution; Springer: Cham, Switzerland; Berlin, Germany , pages 1– 296, 2018

2018
[12]

Ste- gun

Milton Abramowitz and Irene A. Ste- gun. Handbook of Mathematical Func- tions with Formulas, Graphs, and Math- ematical Tables, volume 55. Dover Pub- lications, New York, 1964. Formula 17.4.17. 9

1964

[1] [1]

Nonlinear dynam- ics and chaos: with applications to physics, biology, chemistry, and engi- neering (studies in nonlinearity) , vol- ume 1

Steven H Strogatz. Nonlinear dynam- ics and chaos: with applications to physics, biology, chemistry, and engi- neering (studies in nonlinearity) , vol- ume 1. Westview press, 2001

2001

[2] [2]

Erzwungene Schwingun- gen bei ver¨ anderlicher Eigenfrequenz und ihre technische Bedeutung

Georg Duﬃng. Erzwungene Schwingun- gen bei ver¨ anderlicher Eigenfrequenz und ihre technische Bedeutung . Number 41-

[3] [3]

Non- linear oscillations

Ali H Nayfeh and Dean T Mook. Non- linear oscillations . John Wiley & Sons, 2024

2024

[4] [4]

Lindstedt

A. Lindstedt. Mem. Acad. Imper. Sci. St. Petersburg, 31, 1883

[5] [5]

Ho- motopy analysis approach to duﬃng- harmonic oscillator

Shao-dong Feng and Li-qun Chen. Ho- motopy analysis approach to duﬃng- harmonic oscillator. Applied Mathe- matics and Mechanics , 30(9):1083–1089, 2009

2009

[6] [6]

Applica- tion of he’s energy balance method to duﬃng-harmonic oscillators

M Momeni, N Jamshidi, Amin Barari, and Davood Domiri Ganji. Applica- tion of he’s energy balance method to duﬃng-harmonic oscillators. Interna- tional Journal of Computer Mathemat- ics, 88(1):135–144, 2011

2011

[7] [7]

He’s parameter-expanding methods for strongly nonlinear oscilla- tors

Lan Xu. He’s parameter-expanding methods for strongly nonlinear oscilla- tors. Journal of Computational and Applied Mathematics , 207(1):148–154, 2007

2007

[8] [8]

Perturbation methods

Ali H Nayfeh. Perturbation methods . John Wiley & Sons, 2024

2024

[9] [9]

Non- linear ordinary diﬀerential equations: an introduction for scientists and engineers

Dominic Jordan and Peter Smith. Non- linear ordinary diﬀerential equations: an introduction for scientists and engineers . Oxford University Press, 2007

2007

[10] [10]

A perturbation method for certain non-linear oscillators

TD Burton. A perturbation method for certain non-linear oscillators. Interna- tional Journal of Non-Linear Mechanics , 19(5):397–407, 1984

1984

[11] [11]

Strong nonlinear oscillators

Livija Cveticanin. Strong nonlinear oscillators. Mathematical Engineering. Analytical Solution; Springer: Cham, Switzerland; Berlin, Germany , pages 1– 296, 2018

2018

[12] [12]

Ste- gun

Milton Abramowitz and Irene A. Ste- gun. Handbook of Mathematical Func- tions with Formulas, Graphs, and Math- ematical Tables, volume 55. Dover Pub- lications, New York, 1964. Formula 17.4.17. 9

1964