Computer-Assisted Proofs in Dynamical Systems: A Case Study of a Heteroclinic Orbit in the Shimizu--Morioka System
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: We're looking at the paper: "Computer-Assisted Proofs in Dynamical Systems".
Paper summary: Ada: So, looking at the full paper, the title Computer-Assisted Proofs in Dynamical Systems: A Case Study of a Heteroclinic Orbit in the Shimizu–Morioka System suggests this is more than just a technical exercise. It’s about showing how powerful these computer-assisted proof techniques are when applied to specific problems like finding connecting orbits <ref:2605.07500#pg1>.
Emmy: I think the core idea is that this method, the radii polynomial approach, provides a unified way to validate different dynamical system tasks, from equilibria to invariant manifolds <ref:2605.07500#pg4>. It shows that the underlying mathematical structure is what matters most for these types of proofs one <ref:2605.07500#pg1>.
Kolya: From a probabilistic side, I see it as showing how robust these computational methods can be when we need to check things that are inherently chaotic or highly nonlinear, like in this strange attractor mentioned on page one thirty <ref:2605.07500#pg1>. The constants they find must survive whatever noise or uncertainty is in the real-world application.
Ada: Exactly. It’s about taking a problem where you might get stuck with approximations and giving it a rigorous way to check if those approximations are good enough for the result to hold, like checking that Z is less than one <ref:2605.07500#pg3>.
Emmy: And it connects the abstract mathematical theory with actual code snippets in Julia three, which makes this framework practical for researchers working on these systems today <ref:2605.07500#pg1>. It’s a bridge between theory and computation, which is what this whole paper is about <ref:2605.07500#pg1>.
Kolya: It’s a way to get that kind of validation without having to solve the whole nonlinear differential equation exactly, which is often impossible anyway thirty-four <ref:2605.07500#pg2>. That's a practical implication for anyone trying to understand these systems.
Ada: So it boils down to showing that for this specific system, by following these four steps and checking that Z is less than one, you get a computer-assisted proof of the connecting orbit <ref:2605.07500#pg1>. It's a very concrete way to establish existence where traditional methods might just leave you with an estimate.
Conclusion: Ada: So, we just finished walking through how they used this radii polynomial approach to prove something about a connecting orbit in the Shimizu–Morioka system <ref:2605.07500#pg1>.
Emmy: Yeah, it’s a lot of moving parts for one proof. We’re talking about equilibria, invariant manifolds, and that tricky transverse connection between them.
Kolya: It seems like they used this same four-step process for every single thing they wanted to validate across the whole system. That consistency is what makes it interesting from a structural standpoint <ref:2605.07500#pg4>.
Ada: Exactly, and that structure is what the authors are highlighting—that this method isn't just a trick for this one problem, but a general way to check these kinds of dynamical system things <ref:2605.07500#pg4>.
Emmy: The title itself says "Computer-Assisted Proofs," which tells you the main takeaway is that we can use computation to give us rigorous proof where before we might just have an estimate.
Kolya: It’s about taking something that's hard to prove exactly, like this orbit, and using these computer checks to confirm its existence for sure. That's a big deal in math.
Ada: And it applies this framework to a specific system where the mathematics gets pretty complex with saddle points and flows in three dimensions <ref:2605.07500#pg1>.
Emmy: So, if you only take one thing away from this, it’s that for these kinds of nonlinear problems, we have this reliable template for checking things like equilibria or connections.
Kolya: It gives us a systematic way to look at these hard systems and know exactly what kind of mathematical tool you need to apply first.
Ada: It's adaptable, too, which is the real strength here; they show it works for other systems where the equations are built around this specific structure <ref:2605.07500#pg1>.
Emmy: So, the big implication is that instead of getting stuck on just one system’s weird math, we get a toolkit for checking many different systems with the same method.
Kolya: And it means we can start applying this to other places where we need to verify solutions in complex dynamics.
Ada: The paper sets up this whole framework, and it shows that the computer-assisted proof approach is a solid way forward for tackling these types of dynamical system problems <ref:2605.07500#pg1>.
Olivier Hénot, Akitoshi Takayasu
National Taiwan University · University of Tsukuba
math.DS
Submitted: 2026-05-08
Updated: 2026-11-18
Code: https://github.com/JuliaIntervals/IntervalArithmetic.jl
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The gist: Computer-Assisted Proofs in Dynamical Systems: A Case Study of a Heteroclinic Orbit in the Shimizu–Morioka System is a computer-assisted proof that validates equilibria, invariant manifolds…
Key concepts
- Radii Polynomial Approach
- This is a validation method based on contracting a quasi-Newton operator. It uses an iterative four-step procedure—zero-finding formulation, approximate zero, approximate inverse, and bound estimates—to prove that a solution exists within a specific bounded region.
- Validated Computation of Equilibria
- This involves finding fixed points (equilibria) where the system's function equals zero. The method uses Newton's method to find an approximate zero and then uses interval arithmetic to estimate bounds, ensuring the contraction criterion is met to guarantee a true solution.
- Validated Computation of Invariant Manifolds
- Local invariant manifolds are validated by finding a diffeomorphism that maps the nonlinear flow onto a linear flow. This is done by solving an infinite system of equations using Newton's method on finite approximations to prove the existence and structure of these curves.
- Transverse Intersection
- Proving a connecting orbit exists involves setting up a boundary value problem. The trajectory is expanded using Chebyshev polynomials, and a contraction criterion is used to verify that the intersection between the trajectory and other structures (like saddle equilibria) is transverse.
Terminology
Summary
Computer-Assisted Proofs in Dynamical Systems: A Case Study of a Heteroclinic Orbit in the Shimizu–Morioka System is a computer-assisted proof that validates equilibria, invariant manifolds, and connecting orbits for a transverse heteroclinic orbit in the Shimizu–Morioka system using the radii polynomial approach <ref:2605. p. 007500#pg2>. This method reveals a common structural framework for validating different dynamical systems problems <ref:2605. p. 007500#pg4>.
The Radii Polynomial Approach
The radii polynomial approach is an a posteriori validation method based on the contraction of a quasi-Newton operator <ref:2605.07500#pg2> The procedure involves a four-step process for each subproblem: (i) zero-finding formulation, (ii) approximate zero, (iii) approximate inverse, and (iv) bound estimates. p. 0 Theorem 1.1 provides sufficient conditions to prove the contraction of the quasi-Newton operator x 7→ x−AF(x) in an explicit closed ball of x¯ The interpretation of Y and Z relates to measuring the quality of the approximate zero and its inverse, respectively. p. 2
Validated Computation of Equilibria
The first task involves identifying two distinct equilibria c⋆0, c⋆1 in R3 such that f(c⋆0) = f(c⋆1) = 0 The implementation uses the Radii Polynomial Theorem 1.1 with F = f on X = R3, endowed with the 1-norm. p. 5 Step 2 computes an approximate zero c¯ using Newton’s method, and Step 3 constructs an approximate inverse A ≈ DF(¯c)−1 numerically. p. 4 The final step estimates the bounds Y and Z using interval arithmetic to verify the contraction criterion Z < 1. p. 5
Validated Computation of Invariant Manifolds
The local invariant manifolds are validated via a parameterization method, which seeks a diffeomorphism P defining a topological conjugacy between the nonlinear flow and the linear flow. This involves constructing an infinite system of equations F: XT,ν → XT,ν given by F(P)def = P − ϕ − Diag(LT)f(P). p. 8 The zero-finding problem is solved using Newton’s method on a finite approximation Π≤K ◦ F ◦Π≤K. p. 4
Validated Computation of Transverse Intersection
The existence of the connecting orbit is proven by formulating it as a boundary value problem where the unknowns include local coordinates σ, θ, and integration time τ. The trajectory is expanded in Chebyshev polynomials to define a Banach space XC,µ. p. 8 A zero-finding map F: XC,µ → XC,µ is constructed using the parameterization and the vector field. p. 4 The contraction criterion Z(R) = Z0 + Z1(R) < 1 verifies the transversality of the intersection. p. 5
Conclusion
The analysis successfully provides a computer-assisted proof of a transverse heteroclinic orbit connecting two saddle equilibria in the Shimizu–Morioka system. Each subproblem follows the same four-step template, and the application of the Radii Polynomial Theorem 1. p. 41 closes each argument. p. 3 The approach is adaptable to other systems where F is built around x − ϕ − Lf(x). p. 0
ACKNOWLEDGEMENT
O. Hénot was supported by the National Science and Technology Council (NSTC) under grant No. p. 0 115-2115-M-002-001-MY2. A. Takayasu was supported by the Japan Science and Technology Agency (JST) through the FOREST Program under grant No. p. 0 JPMJFR246A, and JSPS KAKENHI under grant No. 24K00538 and No. 26K00619.
REFERENCES
[1] Àlex Haro, Marta Canadell, Jordi-Lluís Figueras, Alejandro Luque, and Josep Maria Mondelo. The Parameterization Method for Invariant Manifolds: From Rigorous Results to Effective Computations, volume 195 of Applied Mathematical Sciences. p. 0 Springer, Cham, 2016.
[2] Luis Benet, Olivier Hénot, Benoît Richard, and David P. p. 0 Sanders. IntervalArithmetic. p. 1jl.
[3] Jeff Bezanson, Alan Edelman, Stefan Karpinski, and Viral B Shah. Julia: A fresh approach to numerical computing. p. 4
[4] X. Cabré, E. Fontich, and R. de la Llave. The parameterization method for invariant manifolds. p. 0
[5] X. Cabré, E. Fontich, and R. de la Llave. The parameterization method for invariant manifolds II. p. 0
[6] X. Cabré, E. Fontich, and R. de la Llave. The parameterization method for invariant manifolds III. p. 0
[7] Renato Calleja, Carlos García-Azpeitia, Olivier Hénot, Jean-Philippe Lessard, and Jason D Mireles James. p. 0 From the Lagrange triangle to the figure eight choreography: Proof of Marchal’s conjecture. p. 0
[8] Carmen Chicone. Ordinary Differential Equations with Applications, volume 34 of Texts in Applied Mathematics. p. 1 Springer, Cham, 3rd edition, 2024.
[9] Simon Danisch and Julius Krumbiegel. Makie.jl: Flexible high-performance data visualization for Julia. p. 1
[10] Gabriel William Duchesne, Jean-Philippe Lessard, and Akitoshi Takayasu. p. 0 A rigorous integrator and global existence for higher-dimensional semilinear parabolic PDEs via semigroup theory.
[11] Herman H. Goldstine. A History of Numerical Analysis from the 16th Through the 19thCentury, volume 2 of Studies in the History of Mathematics and Physical Sciences. p. 0 SpringerVerlag, New York, 1977. p. 8
[12] Javier Gómez-Serrano. Computer-assisted proofs in PDE: a survey. p. 0
[13] Olivier Hénot. p. 0 RadiiPolynomial. p. 0jl.
[14] Olivier Hénot and Akitoshi Takayasu. p. 0 ShimizuMoriokaTutorial.
[15] Allan Hungria, Jean-Philippe Lessard, and Jason D. Mireles James. Rigorous numerics for analytic solutions of differential equations: the radii polynomial approach. p. 1
[16] Jonathan Jaquette. A proof of Jones’ conjecture. p. 0
[17] Jean-Philippe Lessard, Jason D Mireles James, and Christian Reinhardt. Computerassisted proof of transverse saddle-to-saddle connecting orbits for first order vector fields. p. 0
[18] Edward N. Lorenz. Deterministic nonperiodic flow. p. 6
[19] Mitsuhiro T. Nakao. A numerical approach to the proof of existence of solutions for elliptic problems. p. 1
[20] Mitsuhiro T. Nakao, Michael Plum, and Yoshitaka Watanabe. Numerical verification methods and computer-assisted proofs for partial differential equations, volume 53 of Springer Series in Computational Mathematics. p. 1 Springer, Singapore, [2019] ©2019.
[21] Mitsuhiro T. Nakao and Yoshitaka Watanabe. Learning Verified Numerical Computations through Examples: Theory and Implementation.
[22] Mitsuhiro T. Nakao and Nobito Yamamoto. Verified Numerical Computations — A Computational Challenge to Infinity. p. 1
[23] Shin’ichi Oishi. Numerical verification of existence and inclusion of solutions for nonlinear operator equations. p. 1
[24] Shin’ichi Oishi, editor. Principles of Verified Numerical Computations. p. 2 Corona Publishing, Tokyo, 2018.
[25] Shin’ichi Oishi and Siegfried M. Rump. Fast verification of solutions of matrix equations. p. 1
[26] Christopher Rackauckas and Qing Nie. DifferentialEquations. p. 1jl – a performant and featurerich ecosystem for solving differential equations in Julia. p. 1
[27] Siegfried M. Rump. INTLAB — INTerval LABoratory. p. 2
[28] Siegfried M. Rump. Verification methods: Rigorous results using floating-point arithmetic. p. 3
[30] T. Shimizu and N. p. 0 Morioka. On the bifurcation of a symmetric limit cycle to an asymmetric one in a simple model. p. 6
[31] Teruo Sunaga. Theory of an interval algebra and its application to numerical analysis.
[32] Akitoshi Takayasu, Jean-Philippe Lessard, Jonathan Jaquette, and Hisashi Okamoto. p. 0 Rigorous numerics for nonlinear heat equations in the complex plane of time. p. 1
[33] Lloyd N. Trefethen. Approximation theory and approximation practice. p. 1 Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 2013. p. 1
[34] Warwick Tucker. A rigorous ODE solver and Smale’s 14th problem. p. 1
[35] Minoru Urabe. Galerkin’s procedure for nonlinear periodic systems. p. 0
[36] Jan Bouwe van den Berg, Jason D Mireles James, and Christian Reinhardt. Computing (un)stable manifolds with validated error bounds: Non-resonant and resonant spectra. p. 6
[37] Jan Bouwe van den Berg and Jonathan Jaquette. A proof of Wright’s conjecture. p. 0
[38] Jan Bouwe van den Berg and Jean-Philippe Lessard, editors. Rigorous Numerics in Dynamics, volume 74 of Proceedings of Symposia in Applied Mathematics. p. 1 American Mathematical Society, Providence, RI, 2018. p. 0
[39] Nobito Yamamoto. A numerical verification method for solutions of boundary value problems with local uniqueness by Banach’s fixed-point theorem. p. 0 SIAM Journal on Numerical Analysis, 35(5):2004–2013, 1998. p. 0
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI systems can rigorously validate dynamical system properties like equilibria and invariant manifolds using a four-step procedure that highlights a
unifying structure behind the a posteriori validation method.
This allows for computer-assisted proofs in both finite- and infinite-dimensional settings. -
The system can compute the existence of solutions to boundary value problems, such as transverse heteroclinic orbits, by formulating the problem as a zero-finding map F and applying the Radii Polynomial Theorem 1.1 to bound the error with respect to a norm k · kXT,ν on polynomial approximations.
-
AI systems can compute rigorous error bounds for parameterizations of local invariant manifolds by validating Taylor series expansions using interval arithmetic, yielding an interval of existence whose infimum r is a
rigorous a posteriori error bound.
-
The system can perform the validation of eigenspaces by applying the Radii Polynomial Theorem 1.1 to the eigenvalue problem Df(c⋆i)v = λv, concluding that equilibria are saddle points with specific eigenvalue structures (e.g.,
c⋆0 is a saddle equilibrium with two real stable eigenvalues and one real unstable eigenvalue
). -
AI systems can construct approximate solutions to boundary value problems by performing Newton's method on a finite approximation of the zero-finding problem F, using an initial guess derived from numerical integration and Chebyshev interpolation of the trajectory.
-
The system can compute rigorous bounds for the existence of heteroclinic orbits by solving a boundary value problem formulated as a zero-finding map F on an infinite-dimensional Banach space XC,µ, where the bound Z(R) is calculated via Proposition 4.4 involving terms Z0 and Z1(R).
Related papers
- Proof of the singularity conjecture for discrete subgroups of SL 2(R)
- Multiple ergodic averages for commuting multiplicative actions
- A finite Livshits theorem and local length spectrum rigidity
- Topological Detection of Hopf Bifurcations via Persistent Homology: A Functional Criterion from Time Series
- Rigidity and equidistribution of random walks by diffeomorphisms near the conservative regime
- A hyperbolic area proof of the Kahn-Lyubich Covering Lemma and its generalization