Chaotic dynamics for the two body problem on a sphere

Sergey Bolotin · 2025

Read Chaotic dynamics for the two body problem on a sphere by Sergey Bolotin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-19178.txt.

Introduction

Sergey Bolotin proves that the two-body problem on a sphere admits chaotic trajectories. Two bodies moving under attraction on a curved surface form a classic system, and showing true chaos in it takes care, since most orbits look regular. The trajectories Bolotin constructs pass through near-collisions, echoing the second-species solutions Poincaré found for the classical three-body problem in flat space.

The construction leans on a general result about Lagrangian systems whose potentials have Newtonian singularities, proved by the method of anti-integrable limit developed by Serge Aubry. That method builds complex orbits by gluing simpler pieces in a limiting regime. The result extends the known map of chaos in celestial mechanics from flat space to the sphere, a setting with its own geometry and collision structure.

Catalog record

CREATORSergey Bolotin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-19178

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Chaotic dynamics for the two body problem on a sphere Sergey Bolotin Abstract We prove the existence of chaotic trajectories for the two body problem on a sphere. The trajectories we construct encounter near-collisions and are similar to the second species solutions of Poincaré of the classical 3 body problem. The construction uses a general result on Lagrangian systems with Newtonian singularities of the potential which is based on the method of anti-integrable limit of Serge Aubry. Dedicated to Serge Aubry at his 80th birthday 1 Introduction Celestial mechanics on surfaces of constant curvature (the sphere and the Lobachevsky plane) goes back to the middle of 19th century (see the history in [ 2 , 7 , 10 ] ), and it attracted considerable attention in recent years (see, e.g., [ 7 , 10 , 14 , 18 , 21 , 23 ] ). While it shares many similarities with celestial mechanics in Euclidean space, there are significant differences. For instance, the two-body problem in Euclidean space reduces to the Kepler problem and is therefore integrable. The Kepler problem on surfaces of constant curvature closely resembles its Euclidean counterpart: its bounded trajectories are periodic (Serret, 1890). However, for nonzero curvature, the two-body problem cannot be reduced to the Kepler problem. Numerical evidence (see, e.g., [ 7 ] ) suggests that the two-body problem on a sphere exhibits chaotic behavior in certain regions of phase space, although, to the best of our knowledge, this has never been proven analytically. Several results establish nonintegrability, but only in the meromorphic sense (see e.g. [ 15 , 22 , 24 ] ). However, meromorphic nonintegrability is a much weaker notion than real-analytic nonintegrability and has limited relevance to the system’s real dynamics. Indeed, most linear time periodic systems with meromorphic time dependence are meromorphically non-integrable, yet they are completely integrable in the real sense, possessing a full set of quadratic first integrals. In this paper, we demonstrate chaotic behavior and, in particular, prove real-analytic nonintegrability of the two-body problem on the sphere in the case where one body’s mass is much smaller than the other’s. Then the heavier body moves nearly inertially, and, to first approximation, the motion of the lighter body is described by the restricted two-body problem. The proofs in this paper rely on general results in [ 4 ] concerning Lagrangian systems with Newtonian singularities. The chaotic trajectories we construct pass close to collisions between the bodies. These trajectories resemble the so-called ”second species solutions,” whose existence Poincaré claimed to prove in [ 20 ] for the planar three body problem with masses \varepsilon m_{1},\varepsilon m_{2},1 , where 0<\varepsilon\ll 1 . Second species solutions are those which, in the limit \varepsilon\to 0 , converge to chains of collision orbits – what Poincaré referred to as ”orbites á chocs.” However, his sketch proof of their existence is incomplete, and, as far as we know, rigorous proofs of Poincaré’s claims have never been published. For the planar circular restricted three body problem, numerous results on second species solutions have been obtained by astronomers; see the survey in [ 12 ] . Mathematical contributions include works by Alexeyev [ 1 ] , Perko [ 19 ] , Henrard [ 13 ] , Bruno [ 6 ] , Gomez and Olle [ 11 ] , and Marco and Niederman [ 17 ] . Chaotic second species solutions were constructed in [ 4 ] for the planar circular restricted problem and in [ 5 ] for the spatial restricted three body problem. In [ 16 ] , MacKay and Pinheiro constructed second species solutions describing the motion of charged particles in a magnetic field. 2 Main result Without loss of generality let the radius of the sphere be one, then S^{2}=\{\mathbf{q}\in{\mathbb{R}}^{3}:|\mathbf{q}|=1\} . The Lagrangian of the 2-body problem on the sphere S^{2} is L=\frac{1}{2}m_{1}|\dot{\mathbf{q}}_{1}|^{2}+\frac{1}{2}m_{2}|\dot{\mathbf{q}}_{2}|^{2}+U(\mathbf{q}_{1},\mathbf{q}_{2}), where U is the interaction potential. The Hamiltonian on T^{*}(S^{2}\times S^{2}) is H=\frac{1}{2m_{1}}|\mathbf{p}_{1}|^{2}+\frac{1}{2m_{2}}|\mathbf{p}_{2}|^{2}-U(\mathbf{q}_{1},\mathbf{q}_{2}), with the standard symplectic structure on T^{*}S^{2}=\{(\mathbf{q},\mathbf{p}):|\mathbf{q}|=1,\langle\mathbf{p},\mathbf{q}\rangle=0\}. The natural analog of the Newtonian gravitational potential for the sphere is U(\mathbf{q}_{1},\mathbf{q}_{2})=\frac{k}{\tan\rho},\qquad\rho=d(\mathbf{q}_{1},\mathbf{q}_{2}), where the distance is with respect to the standard metric on S^{2} : \rho=\arccos\langle\mathbf{q}_{1},\mathbf{q}_{2}\rangle. Explicitly U(\mathbf{q}_{1},\mathbf{q}_{2})=\frac{k\langle\mathbf{q}_{1},\mathbf{q}_{2}\rangle}{\sqrt{1-\langle\mathbf{q}_{1},\mathbf{q}_{2}\rangle^{2}}}. (1) The potential is singular on \Delta=\{(\mathbf{q}_{1},\mathbf{q}_{2})\in S^{2}\times S^{2}:\mathbf{q}_{1}=\pm\mathbf{q}_{2}\} . As \rho\to 0 , we have U\sim\frac{k}{\rho}, (2) as for the Newtonian potential. Usually one sets k=\gamma m_{1}m_{2} . There are two reasons why the potential ( 1 ) is the natural analog of the Newtonian potential: • For fixed \mathbf{a}\in S^{3} , the function V(\mathbf{q})=U(\mathbf{q},\mathbf{a}) on the 3-dimensional sphere satisfies the Laplace-Beltrami equation, just as the Newtonian potential on {\mathbb{R}}^{3} (Schering 1873). • For the Kepler problem on a sphere (when one of the particles is fixed), the motion of the second particle in the potential field V(\mathbf{q}) is periodic (Serret 1890), exactly as for the Kepler problem in the Euclidean space. Darboux (1884) proved a generalization of Bertrand’s theorem: the only other such potential is the analog k\tan^{2}\rho of the Hooke potential. However, for the results below it is only important that U is SO(3) invariant and has asymptotics ( 2 ) as \rho\to 0 . Contrary to the 2-body problem in {\mathbb{R}}^{2} , the 2-body problem on S^{2} is not integrable: the integrals of energy H and the angular momentum \mathbf{M}=\mathbf{q}_{1}\times\mathbf{p}_{1}+\mathbf{q}_{2}\times\mathbf{p}_{2} corresponding to the SO(3) symmetry are insufficient for integrability. For fixed angular momentum \mathbf{M}=\mathbf{M}_{0}\neq 0 , only the SO(2) symmetry of rotation around \mathbf{M}_{0} remains. Hence the reduced Hamiltonian system has two degrees of freedom. Explicit reduction was first performed by Shchepetilov [ 21 , 22 ] , and then a simpler reduction was done in [ 7 , 9 ] . Numerical evidence (see e.g. [ 7 ] ) shows that the 2-body problem on S^{2} is chaotic in certain regions of the phase space but, as far as we know, this was never proved analytically. There are several results establishing nonintegrability, but only in the meromorphic sense [ 15 , 22 , 24 ] which has limited relation to real dynamics of the system. We establish chaotic behavior and, in particular, real-analytic nonintegrability of the two body problem on S^{2} in the case when one mass is much smaller than the other. Theorem 1 . Fix constants -1/2<a<b and let \varepsilon_{0}>0 and \sigma_{0}>0 be sufficiently small. If m_{2}/m_{1}<\sigma_{0} , the 2-body problem on S^{2} has a chaotic invariant set on the level sets \{H=h,\mathbf{M}=\mathbf{M}_{0}\} of the first integrals for a<\frac{m_{1}^{2}h}{m_{2}M_{0}^{2}}<b,\quad\frac{m_{1}^{2}k}{m_{2}M_{0}^{2}}<\varepsilon_{0}. In many respects celestial mechanics on a sphere is similar to celestial mechanics on a Lobachevsky space, see e.g. [ 9 , 10 ] . In particular, meromorphic nonintegrability still holds, see [ 15 , 22 ] . But an analog of Theorem 1 will not hold: for similar values of the first integrals trajectories will escape to infinity. As mentioned in the introduction, chaotic trajectories whose existence is claimed in Theorem 1 encounter near collisions and are similar in nature to the “second species solutions” of Poincaré [ 20 ] of the Euclidean 3 body problem. 3 Restricted 2 body problem The proof of Theorem 1 is based on the reduction to the restricted 2 body problem on S^{2} . We recall a formal derivation. In the formal limit m_{1}\to+\infty , the light particle does not

Read the complete text