betaFeedback

WorksMathematics

Collision Systems, with Applications to Geodesic Networks and Point Billiards

Abstract

A collision system is a finite set of point particles with masses, moving freely with velocities given by the gradient of a convex energy function of the momentum, whose collisions conserve total momentum and total energy and are otherwise unconstrained. Stationary geodesic networks near a line of multiplicity nn, point billiards with momentum- and energy-conserving collisions, and relativistic point particles are collision systems. We prove that the number of collisions of a collision system is at most exponential in its total mass, with an explicit base, whenever the energy is close enough to the Newtonian one, and that it can be exponential for every energy on a line. For geodesic networks this replaces the previously reported bounds, of order n2n^2 from below and nn2n^{n^2} from above, by exponential bounds on both sides; for point billiards it gives explicit exponential bounds. The exponential growth rates remain open.

Introduction

How many times can nn particles collide? For hard balls in Euclidean space the question was raised by Sinai, shown to have a finite answer by Vaserstein [18] and Gal’perin [11], and made quantitative by Burago, Ferleger and Kononenko [2], whose bound (32n3/2)n2(32n^{3/2})^{n^2} for equal balls was improved to nOd(n)n^{O_d(n)} in each fixed dimension dd by Burdzy [5]; Burago and Ivanov [3] constructed examples with 2⌊n/2⌋2^{\lfloor n/2\rfloor} collisions in dimension three. The same question arises in two other settings, with the same shape of answer so far. The first is geometric: an integral stationary geodesic network in Rd\mathbb{R}^d that converges weakly to a line of multiplicity nn has, on a compact set, a bounded number of singular points, by the partial regularity theorem of Bosch [4]; that paper reports the estimate (10n)n2(10n)^{n^2} in the plane, while nn lines in general position give (n2)\binom{n}{2} singular points, and nothing better was known on either side. The second is the linear point-billiard framework of Féjoz, Knauf and Montgomery [9], for which a uniform finite collision bound is known. Gasiorek [10] gives an explicit bound for three equal-mass particles in the plane. Here we allow any momentum- and energy-conserving outcome for an entire coincident cluster, including collisions of three or more particles and flights of coincident particles.

This paper treats these problems in one framework. A collision system is a finite set of point particles with masses, moving in a Euclidean space with velocities grad⁡h(pi)\operatorname{grad} h(p_i) given by a convex energy hh of the momentum, whose collisions conserve total momentum ∑mipi\sum m_i p_i and total energy ∑mih(pi)\sum m_i h(p_i) and are otherwise unconstrained. The particles may have arbitrary positive masses, and our bounds specialize to identical particles. The Newtonian energy 12∣p∣2\frac{1}{2}|p|^2 gives point billiards. The framework applies to geometry because a stationary geodesic network near a line of multiplicity nn is a collision system too: reading the axial coordinate as time, stationarity at a vertex is exactly the conservation of the transverse part pp of the unit tangent and of its longitudinal defect 1−1−∣p∣21-\sqrt{1-|p|^2}, so the nn strands are particles with the energy h(p)=1−1−∣p∣2h(p)=1-\sqrt{1-|p|^2}, which is Newtonian to second order. Relativistic point particles, with the energy 1+∣p∣2\sqrt{1+|p|^2}, are a third instance.

Two parameters appear in the bounds, and they should be distinguished. The number nn of particles is the natural parameter when all particles are alike, and it is the parameter in the questions above. For particles of different masses the right parameter is the ratio of the total mass to the smallest mass, which we call MM. Multiplying all masses by a constant changes nothing, so we normalize the smallest mass to 11; then every mass is at least 11 and M=m1+⋯+mn≥nM=m_1+\cdots+m_n\ge n, with equality exactly when all masses are equal. Thus M=nM=n for nn particles of unit mass, for nn strands of weight one and for nn balls of equal mass, and a bound in terms of MM is a bound in terms of nn for them. Our upper bounds are proved for arbitrary masses and are stated in terms of MM; our lower bounds are constructions with unit masses and are stated in terms of nn.

We prove two theorems about collision systems, stated in § 1.2: an exponential upper bound M28M−2\frac{M}{2}8^{M-2} on the number of collisions of a system of total mass MM, in any dimension, for energies close enough to the Newtonian one, closeness being measured by the pinching of hh, a uniform bound on its normalized Hessian; and an exponential lower bound, a single construction that produces, for every energy on a line and every n≥5n\ge5, systems of nn particles of unit mass with at least 2⌊(n−2)/3⌋2^{\lfloor(n-2)/3\rfloor} collisions at which at least three particles meet. Each application is then a verification that the system in question is a collision system with a specific energy. Table 1 summarizes what the framework yields against what was known.

ProblemKnownThis paper
Geodesic networks C(d,n)C(d,n)upperd3(n−1)/2(5n)n2d^{3(n-1)/2}(5n)^{n^2} [4]n28n−2\frac{n}{2}8^{n-2}
lower(n2)\binom{n}{2}2⌊(n−2)/3⌋2\lfloor(n-2)/3\rfloor
Point billiards, total mass MMupperfinite [9, 2]M28M−2\frac{M}{2}8^{M-2}
lower(n2)\binom{n}{2}2⌊(n−2)/3⌋2\lfloor(n-2)/3\rfloor
Relativistic particles, slowupper—M28M−2\frac{M}{2}8^{M-2}

Table 1. Bounds on the number of collisions (singular points, for networks); Theorems A, B and C. Lower bounds are for n≥2n \ge2, with equal masses, and are realised in the plane, respectively on a line; upper bounds hold in every dimension. The earlier network estimate was reported in [4]; point-billiard finiteness refers to the conventions of [9], and an explicit planar three-particle bound appears in [10].

For geodesic networks the gap between n2n^2 and nn2n^{n^2} is thus closed to a gap between two exponentials, 2n/32^{n/3} and 8n8^n, and the upper bound no longer depends on the dimension. For point billiards we obtain bounds for arbitrary particle counts under the cluster collision rule above. The exponential growth rates remain open, as does their dependence on the dimension and the optimal dependence on the mass ratio for a fixed number of particles.

Uniform upper bounds outside the near-Newtonian regime are not addressed here.

The paper is organized as follows. The rest of this introduction defines collision systems, states the two general theorems and the three applications, and describes the method. § 2 contains the examples of energies and a rigidity proposition for systems in which all particles meet at one point, § 3 the proof of the exponential upper bound, § 4 the construction for the lower bound, and § 5 the applications. Three longer computations are in the appendices.

Collision systems

Definition 1.1. An energy on XX is a pair (Ω,h)(\Omega,h), where Ω⊂X\Omega\subset X is open, convex and contains 00, and h:Ω→Rh:\Omega\to\mathbb{R} is smooth with

h(0)=0,grad⁡h(0)=0,D2h(p)≥I(p∈Ω),h(0)=0,\qquad\operatorname{grad} h(0)=0,\qquad D^2h(p)\ge I\quad(p\in\Omega),

and such that the set of velocities V=grad⁡h(Ω)V=\operatorname{grad} h(\Omega) is convex. Its pinching is the supremum ρ(h)∈[1,∞]\rho(h)\in[1,\infty] of the largest eigenvalue of D2h(p)D^2h(p) over p∈Ωp\in\Omega. When Ω\Omega is clear from the context, we write hh for the energy. A point p∈Ωp\in\Omega is a momentum; grad⁡h(p)\operatorname{grad} h(p) is its velocity and h(p)h(p) its energy. Since D2h≥ID^2h\ge I, the map grad⁡h:Ω→V\operatorname{grad} h:\Omega\to V is a diffeomorphism onto the open set VV.

Subtracting a constant and a linear function normalizes h(0)h(0) and grad⁡h(0)\operatorname{grad} h(0); the linear term amounts to a Galilean change of frame. If the Hessian has a positive uniform lower bound aIaI, multiplying hh by 1/a1/a gives D2h≥ID^2h\ge I and amounts to a change of the unit of time. When aa is the optimal lower bound, the pinching is the ratio of the uniform upper and lower Hessian bounds. This is a global bound over Ω\Omega, not the pointwise condition number. Strict convexity alone need not provide a positive uniform lower bound on the whole domain. The Newtonian energy h(p)=12∥p∥2h(p)=\frac{1}{2}\lVert p\rVert^2 on Ω=X\Omega= X has pinching 1, and an energy of pinching close to 1 is an approximation of it.

Definition 1.2. Let (Ω,h)(\Omega,h) be an energy on XX, let m1,…,mn>0m_1,\ldots,m_n>0 be masses, and let J⊂RJ\subset\mathbb{R} be an interval. A collision system with these masses and energy hh on JJ is a list of continuous piecewise affine functions y1,…,yn:J→Xy_1,\ldots,y_n:J\to X, each with finitely many break points, such that:

(a) on every open interval of JJ that contains no break point of any yiy_i (a flight) there are momenta p1,…,pn∈Ωp_1,\ldots,p_n\in\Omega, constant on the flight, with yi′=grad⁡h(pi)y_i'=\operatorname{grad}h(p_i);

(b) for every t∈Jt\in J and y∈Xy\in X, the set S={i:yi(t)=y}S=\{i:y_i(t)=y\} satisfies

∑i∈Smipi(t+)=∑i∈Smipi(t−),∑i∈Smih(pi(t+))=∑i∈Smih(pi(t−)).\sum_{i\in S}m_i p_i(t+)=\sum_{i\in S}m_i p_i(t-),\qquad \sum_{i\in S}m_i h(p_i(t+))=\sum_{i\in S}m_i h(p_i(t-)).

where pi(t±)p_i(t_\pm) are the momenta on the flights just after and just before tt (at an endpoint of JJ the condition is void).

The index ii is the label of the particle yiy_i, mipim_i p_i is its momentum and mih(pi)m_i h(p_i) its energy, and m1+⋯+mnm_1+\cdots+m_n is the total mass. If ∣S∣≥2|S|\geq2 and the functions yiy_i, i∈Si\in S, do not all coincide on a neighbourhood of tt in JJ, the point (t,y)(t,y) is a collision point; a time at which some momentum changes is an event time.

Multiplying all masses by a positive constant changes nothing, so we normalize the smallest mass to 1; the total mass is then the sum of the masses divided by the smallest one. At a collision of two particles of equal mass on a line, the unordered pair of momenta is preserved. A collision of three or more particles, or of two particles in higher dimension, is a relation rather than a map: every outcome compatible with the two conservation laws is allowed, and coincident particles may travel together. For the Newtonian energy, this extends the linear point-billiard framework of Féjoz, Knauf and Montgomery [9]; their stated conventions exclude flights inside collision subspaces and select one collision subspace at a multiple meeting.

The general theorems

Theorem 1.3 (Upper bound). For every M≥1M\geq1 there is ρM>1\rho_M>1, nonincreasing in MM, such that every collision system, on any interval, with masses at least 1, total mass MM, and an energy of pinching less than ρM\rho_M, has at most ⌊M28M−2⌋\left\lfloor\frac{M}{2}8^{M-2}\right\rfloor collision points.

For an integer n≥1n\geq1 put

Un=⌊n28n−2⌋,U_n=\left\lfloor\frac{n}{2}8^{n-2}\right\rfloor,

so that U1=0U_1=0 and U2=1U_2=1. The Newtonian energy has pinching 1, so a collision system with masses at least 1, total mass MM and the Newtonian energy has at most ⌊M28M−2⌋\left\lfloor\frac{M}{2}8^{M-2}\right\rfloor collision points, with no further hypothesis.

References

  1. [1]W. K. Allard and F. J. Almgren, Jr., The structure of stationary one dimensional varifolds with positive density, Invent. Math. 34 (1976), 83–97.DOI
  2. [2]D. Burago, S. Ferleger, and A. Kononenko, Uniform estimates on the number of collisions in semi-dispersing billiards, Ann. of Math. (2) 147 (1998), 695–708.DOI
  3. [3]D. Burago and S. Ivanov, Examples of exponentially many collisions in a hard ball system, Ergodic Theory Dynam. Systems 41 (2021), no. 9, 2754–2769; arXiv:1809.02800.DOI
  4. [4]H. Bosch, A partial e-regularity theorem for integral stationary geodesic networks in R^n, arXiv:2607.23872v1 (2026).arxiv.org/abs/2607.23872
  5. [5]K. Burdzy, An improved upper bound on the number of billiard ball collisions, Comm. Math. Phys. 391 (2022), no. 1, 107–117; arXiv:2107.10362.DOI
  6. [6]O. Chodosh and C. Mantoulidis, The p-widths of a surface, Publ. Math. Inst. Hautes Études Sci. 137 (2023), 245–342.DOI
  7. [7]J. Liu, G. Gao, Z. Sun, B. Wu, S. Wu, S. Liu, J. Jiang, H. Ju, L. Chen, R. Cheng, X. Zhang, and B. Dong, Danus: Orchestrating mathematical reasoning agents with fact-graph memory, arXiv:2607.06447 (2026).arxiv.org/abs/2607.06447
  8. [8]I. Frolov, D. Gorodkov, D. Korshunov, and A. Nabutovsky, Geodesic nets on the Euclidean plane and closed geodesic nets on Riemannian surfaces, arXiv:2606.14080v2 (2026).arxiv.org/abs/2606.14080
  9. [9]J. Féjoz, A. Knauf, and R. Montgomery, Lagrangian relations and linear point billiards, Nonlinearity 30 (2017), no. 4, 1326–1355.
  10. [10]S. Gasiorek, Counting collisions in an N-billiard system using angles between collision subspaces, SIGMA 16 (2020), 119, 13 pp., doi:10.3842/SIGMA.2020.119.DOI
  11. [11]G. A. Gal’perin, Systems of locally interacting and repelling particles that are moving in space (Russian), Trudy Moskov. Mat. Obshch. 43 (1981), 142–196.
  12. [12]G. Galperin, Playing pool with π (the number π from a billiard point of view), Regul. Chaotic Dyn. 8 (2003), 375–394.
  13. [13]A. Nabutovsky and F. Parsch, Geodesic nets: some examples and open problems, arXiv:1904.00483 (2019).arxiv.org/abs/1904.00483
  14. [14]F. Parsch and H. Zhang, Towards constructing geodesic nets with four boundary ver-tices and an increasing number of balanced vertices, arXiv:2511.07810 (2025).arxiv.org/abs/2511.07810
  15. [15]H. Qiu, Example of exponentially many collisions of identical hard disks in the plane, arXiv:2610.14282 (2026).arxiv.org/abs/2610.14282
  16. [16]Ya. G. Sinai, Billiard trajectories in a polyhedral angle (Russian), Uspekhi Mat. Nauk 33 (1978), no. 1(199), 229–230.
  17. [17]S. Tabachnikov, Geometry and billiards, Student Mathematical Library 30, Amer. Math. Soc., Providence, RI, 2005.DOI
  18. [18]L. N. Vaserstein, On systems of particles with finite-range and/or repulsive interactions, Comm. Math. Phys. 69 (1979), 31–56.DOI

Paper details

vlnc:W_bcuqmwzq
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

OpenAI Codex, OpenAI ChatGPT, Frenzymath Danus, Anthropic Claude Fable, Anthropic Claude Opus

View statement
Looking for related papers…
Contents