The N N -body problem on coadjoint orbits
Holger Dullin & Richard Montgomery · 2025
Read The N N -body problem on coadjoint orbits by Holger Dullin & Richard Montgomery (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-2504-01272.txt.
Introduction
Holger Dullin and Richard Montgomery recast in 2025 the classic problem of gravitating bodies in a geometric language. After fixing the center of mass and reducing by rotations, the spatial N-body problem at nonzero spin becomes a singular space with two layers. One covers full spatial motion, the other flat planar motion, and the second forms the frontier of the first.
Each layer turns out to be a coadjoint orbit of a symplectic group, a tidy algebraic home. The result is a special case of a theorem for any dimension, proved with a Poisson reduction tool. The paper also shows Newton's equations can be rewritten as a Lax pair, matching an older matrix form.
Catalog record
| CREATOR | Holger Dullin & Richard Montgomery |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 77 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-01272
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The N N -body problem on coadjoint orbits Holger Dullin & Richard Montgomery Abstract . We show (Theorem 3) that the symplectic reduction of the spatial n n -body problem at non-zero angular momentum is a singular symplectic space consisting of two symplectic strata, one for spatial motions and the other for planar motions. Each stratum is realized as coadjoint orbit in the dual of the Lie algebra of the linear symplectic group S p ( 2 n − 2 ) Sp(2n-2) . The planar stratum arises as the frontier upon taking the closure of the spatial stratum. We reduce by going to center-of-mass coordinates to reduce by translations and boosts and then performing symplectic reduction with respect to the orthogonal group O ( 3 ) O(3) . The theorem is a special case of a general theorem (Theorem 2) which holds for the n n -body problem in any dimension d d . This theorem follows largely from a “Poisson reduction” theorem, Theorem 1. We achieve our reduction theorems by combining the Howe dual pair perspective of reduction espoued in [ 11 ] with a normal form arising from a symplectic singular value decomposition due to Xu [ 18 ] . We begin the paper by showing how Poisson reduction by the Galilean group rewrites Newton’s equations for the n n -body problem as a Lax pair. In section 6.4 we show that this Lax pair representation of the n n -body equations is equivalent to the Albouy-Chenciner [ 2 ] representation in terms of symmetric matrices. 1. Introduction Newton’s n -body problem has the Galilean group as a symmetry group. As a result we can push the differential equations defining the problem down to reduced equations on a kind of quotient space of n -body phase space by the Galilean group. We show here that these reduced equations form a Lax pair: (1) \dot{K}=[P,K],\qquad K,P\in\mathfrak{g}=sp(2n-2). and that these in turn are equivalent to the Albouy-Chenciner version of the reduced N-body equations described in their celebrated paper [ 2 ] . The evolving matrix K(t)\in\mathfrak{g}=sp(2n-2) of ( 1 ) encodes the “Galilean shape” of a moving phase point. The Lie algebra \mathfrak{g} consists of 2n-2\times 2n-2 real square matrices K satisfying KJ+JK^{t}=0 where J is the constant (2n-2)\times(2n-2) matrix representing the standard symplectic form on \mathbb{R}^{2n-2} and described in equation ( 9 ) below. The quotient space is a conjugation invariant convex subset of \mathfrak{g} . The matrix function P=P(K)\in\mathfrak{g} occuring in ( 1 ) corresponds to the Hamiltonian vector field. P is given by equation ( 9 ) below and depends on the masses of the bodies and the potential which defines the problem, assumed to be a function of the mutual distances between bodies alone. Duality. It will be essential to view the Lax pair equations ( 1 ) as being a Lie-Poisson equations on the dual space \mathfrak{g}^{*} of \mathfrak{g} . Being a simple Lie algebra, \mathfrak{g} is isomorphic to \mathfrak{g}^{*} as a G=Sp(2n-2) -space. This isomorphism is unique up to scale and we call it the Killing isomorphism. The isomorphism takes any Lax pair equation on \mathfrak{g} to a Lie-Poisson equation on \mathfrak{g}^{*} . The former equations automatically have the adjoint orbits as invariant submanifolds while the latter set of equations have the corresponding image coadjoint orbits of \mathfrak{g}^{*} as invariant symplectic submanifolds. Symplectic Reduction and New Results. Our main result, theorem 2 , relates certain singular symplectic spaces consisting of the closure of certain coadjoint orbits in \mathfrak{g}^{*} to the (Marsden-Weinstein-Meyer) symplectic reduced spaces for the n -body problem, reduced at certain values of angular momentum. In particular, we show in theorem 3 that the symplectic reduction of the spatial n -body problem with respect to the orthogonal group O(3) at nonzero angular momentum consists of a singular space which is the union of two \mathfrak{g}^{*} coadjoint orbits. We obtain these results by combining the dual pair perspective of [ 11 ] (sections 4 and 5 therein), with a matrix normal form due to Xu [ 18 ] . We summarize the dual pair perspective by diagram ( 2 ) below. We describe the matrix normal form of Xu in Appendix B. 1.1. Forming the quotient. Newton’s equations for n bodies moving in Euclidean space \mathbb{R}^{d} form a set of d\times n second order differential equations which we can rewrite as a system of first order Hamiltonian equations on a phase space P_{d,n} of dimension 2nd . To get from (q,p)\in P_{d,n} to the matrix K of ( 1 ) follow the diagram down to G(Z)=Z^{t}Z and set K=JG . (2) \textstyle{P_{d,n}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{\pi_{trans}} \textstyle{P_{d,n-1}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{L} \scriptstyle{G} \textstyle{o(d)^{*}} \textstyle{sp(2n-2)^{*}} \textstyle{(q,p)\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{\pi_{trans}} \textstyle{Z=(X,Y)\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{L} \scriptstyle{G} \textstyle{ZJZ^{t}} \textstyle{Z^{t}Z} The intermediate object Z\in P_{d,n-1} is a d\times 2n-2 matrix representing the phase point (q,p)\in P_{d,n} after quotient by Galilean boosts and translations. We implement this quotient by the usual trick of going into center-of-mass frame and then using Jacobi vectors. The quotient map \pi_{trans}:P_{d,n}\to P_{d,n-1} is a surjective linear map which implements the quotient. See section 6.2 for details. Remark 1 . P_{d,n-1} is not literally a quotient by Galilean boosts and translations because the boosts do not act on phase space in the usual sense. Boosts act on Galilean space-time and their action requires explicit knowledge of the current time, while the phase spaces keep no knowledge of clock time. Nevertheless, the usual center-of-mass trick allows us to non-ambiguously identify P_{d,n-1} with the quotient of P_{d,n} by the action of the Galilean boosts and translations of \mathbb{R}^{d} . The orthogonal group O(d) is the subgroup of the Galilean group which fixes the center of mass 0\in\mathbb{R}^{d} and acts on centered phase space P_{d,n-1} , with g\in O(d) acting by Z\mapsto gZ . The Gram matrix (3) G=Z^{t}Z\in symm(2n-2) is invariant under this action, and is a complete invariant. Here symm(2n-2) denotes the vector space of all symmetric (2n-2)\times(2n-2) matrices. The image of the Gram map represents the full quotient of P_{d,n} by the Galilean group. See the end of section 4 and section 6.3 for more. We need to explain how we think of G=G(Z) as an element of \mathfrak{g}=sp(2n-2) and of \mathfrak{g}^{*}=sp(2n-2)^{*} . Multiply G by the previously mentioned symplectic structure J to arrive at (4) symm(2n-2)\cong sp(2n-2)\qquad;\qquad G\to JG:=K\in\mathfrak{g}. The matrix K=JG=JG(\pi_{trans}(q,p)) is the matrix appearing in equation ( 1 ). And if q(t),p(t) satisfies Newton’s equations then K(t) satisfies ( 1 ). When we want to view G or K as lying in \mathfrak{g}^{*} use the Killing isomorphism \mathfrak{g}\to\mathfrak{g}^{*} mentioned earlier. Matrices of the form ( 3 ) are not arbitrary symmetric matrices, rather they are positive semi-definite. Moreover, rank(G)\leq d since rank(Z)\leq d . Write symm_{+}(2n-2;d) for the subset of symm(2n-2) consisting of positive semi-definite matrices of rank d or less. Then (5) im(G)=symm_{+}(2n-2;d)\subset symm(2n-2)\cong\mathfrak{g}^{*} is the image of G . The Lie group Sp(2n-2) whose Lie algebra is \mathfrak{g} acts on P_{d,n-1} by right multipication Z\mapsto ZT,T\in Sp(2n-2) . Essential to what follows is that the Gram map is the momentum map for this action after making the identifications symm(2n-2)\cong sp(2n-2)^{*} just described. As a consequence G is a Poisson map. Putting this together with some understanding of