An introduction to the classical three-body problem: From periodic solutions to instabilities and chaos
AAn introduction to the classical three-body problem
From periodic solutions to instabilities and chaos G OVIND
S. K
RISHNASWAMI AND H IMALAYA S ENAPATI
Chennai Mathematical Institute, SIPCOT IT Park, Siruseri 603103, IndiaEmail: [email protected], [email protected]
January 22, 2019Published in Resonance (1), 87-114, January (2019) Abstract
The classical three-body problem arose in an attempt to understand the effect of the Sunon the Moon’s Keplerian orbit around the Earth. It has attracted the attention of some of thebest physicists and mathematicians and led to the discovery of chaos. We survey the three-body problem in its historical context and use it to introduce several ideas and techniquesthat have been developed to understand classical mechanical systems.
Keywords : Kepler problem, three-body problem, celestial mechanics, classical dynamics, chaos, instabilities
Contents a r X i v : . [ n li n . C D ] J a n Introduction
The three-body problem is one of the oldest problems in classical dynamics that continues to throwup surprises. It has challenged scientists from Newton’s time to the present. It arose in an attempt tounderstand the Sun’s effect on the motion of the Moon around the Earth. This was of much practicalimportance in marine navigation, where lunar tables were necessary to accurately determine longitude atsea (see Box 1). The study of the three-body problem led to the discovery of the planet Neptune (see Box2), it explains the location and stability of the Trojan asteroids and has furthered our understanding of thestability of the solar system [1]. Quantum mechanical variants of the three-body problem are relevant tothe Helium atom and water molecule [2].
Box 1:
The
Longitude Act (1714) of the British Parliament offered £ a , Alexis Clairaut and Jean-Baptiste d’Alembert competed todevelop a theory accounting for solar perturbations to the motion of the Moon around the Earth. Fora delightful account of this chapter in the history of the three-body problem, including Clairaut’sexplanation of the annual ◦ rotation of the lunar perigee (which had eluded Newton), see [11].Interestingly, Clairaut’s use of Fourier series in the three-body problem (1754) predates their use byJoseph Fourier in the analysis of heat conduction! a Euler had gone blind when he developed much of his lunar theory!
Box 2: Discovery of Neptune:
The French mathematical astronomer Urbain Le Verrier (1846)was intrigued by discrepancies between the observed and Keplerian orbits of Mercury and Uranus.He predicted the existence of Neptune (as was widely suspected) and calculated its expected po-sition based on its effects on the motion of Uranus around the Sun (the existence and location ofNeptune was independently inferred by John Adams in Britain). The German astronomer JohannGalle (working with his graduate student Heinrich d’Arrest) discovered Neptune within a degreeof Le Verrier’s predicted position on the very night that he received the latter’s letter. It turned outthat both Adams’ and Le Verrier’s heroic calculations were based on incorrect assumptions aboutNeptune, they were extremely lucky to stumble upon the correct location!The three-body problem admits many ‘regular’ solutions such as the collinear and equilateral periodicsolutions of Euler and Lagrange as well as the more recently discovered figure-8 solution. On the otherhand, it can also display chaos as serendipitously discovered by Poincar´e. Though a general solution inclosed form is not known, Sundman while studying binary collisions, discovered an exceptionally slowlyconverging series representation of solutions in fractional powers of time.The importance of the three-body problem goes beyond its application to the motion of celestialbodies. As we will see, attempts to understand its dynamics have led to the discovery of many phenomena(e.g., abundance of periodic motions, resonances (see Box 3), homoclinic points, collisional and non-collisional singularities, chaos and KAM tori) and techniques (e.g., Fourier series, perturbation theory,canonical transformations and regularization of singularities) with applications across the sciences. Thethree-body problem provides a context in which to study the development of classical dynamics as wellas a window into several areas of mathematics (geometry, calculus and dynamical systems).2 ox 3: Orbital resonances:
The simplest example of an orbital resonance occurs when the periodsof two orbiting bodies (e.g., Jupiter and Saturn around the Sun) are in a ratio of small whole numbers( T S /T J ≈ / ). Resonances can enhance their gravitational interaction and have both stabilizingand destabilizing effects. For instance, the moons Ganymede, Europa and Io are in a stable orbital resonance around Jupiter. The Kirkwood gaps in the asteroid belt are probably due to thedestabilizing resonances with Jupiter. Resonances among the natural frequencies of a system (e.g.,Keplerian orbits of a pair of moons of a planet) often lead to difficulties in naive estimates of theeffect of a perturbation (say of the moons on each other). As preparation for the three-body problem, we begin by reviewing some key features of the two-bodyproblem. If we ignore the non-zero size of celestial bodies, Newton’s second law for the motion of twogravitating masses states that m ¨ r = α ( r − r ) | r − r | and m ¨ r = α ( r − r ) | r − r | . (1)Here, α = Gm m measures the strength of the gravitational attraction and dots denote time deriva-tives. This system has six degrees of freedom, say the three Cartesian coordinates of each mass r =( x , y , z ) and r = ( x , y , z ) . Thus, we have a system of 6 nonlinear (due to division by | r − r | ),second-order ordinary differential equations (ODEs) for the positions of the two masses. It is convenientto switch from r and r to the center of mass (CM) and relative coordinates R = m r + m r m + m and r = r − r . (2)In terms of these, the equations of motion become M ¨ R = 0 and m ¨ r = − α | r | r . (3)Here, M = m + m is the total mass and m = m m /M the ‘reduced’ mass. An advantage of thesevariables is that in the absence of external forces the CM moves at constant velocity, which can be chosento vanish by going to a frame moving with the CM. The motion of the relative coordinate r decouplesfrom that of R and describes a system with three degrees of freedom r = ( x, y, z ) . Expressing theconservative gravitational force in terms of the gravitational potential V = − α/ | r | , the equation for therelative coordinate r becomes ˙ p ≡ m ¨ r = −∇ r V = − (cid:18) ∂V∂x , ∂V∂y , ∂V∂z (cid:19) (4)where p = m ˙ r is the relative momentum. Taking the dot product with the ‘integrating factor’ ˙ r =( ˙ x, ˙ y, ˙ z ) , we get m ˙ r · ¨ r = ddt (cid:18) m ˙ r (cid:19) = − (cid:18) ∂V∂x ˙ x + ∂V∂y ˙ y + ∂V∂z ˙ z (cid:19) = − dVdt , (5)which implies that the energy E ≡ m ˙ r + V or Hamiltonian p m + V is conserved. The relativeangular momentum L = r × m ˙ r = r × p is another constant of motion as the force is central : The conservation of angular momentum in a central force is a consequence of rotation invariance: V = V ( | r | ) is indepen-dent of polar and azimuthal angles. More generally, Noether’s theorem relates continuous symmetries to conserved quantities. L = ˙ r × p + r × ˙ p = 0 + 0 . The constancy of the direction of L implies planar motion in the CMframe: r and p always lie in the ‘ecliptic plane’ perpendicular to L , which we take to be the x - y plane with origin at the CM (see Fig. 1). The Kepler problem is most easily analyzed in plane-polarcoordinates r = ( r, θ ) in which the energy E = m ˙ r + V eff ( r ) is the sum of a radial kinetic energyand an effective potential energy V eff = L z mr + V ( r ) . Here, L z = mr ˙ θ is the vertical componentof angular momentum and the first term in V eff is the centrifugal ‘angular momentum barrier’. Since L (and therefore L z ) is conserved, V eff depends only on r . Thus, θ does not appear in the Hamiltonian: itis a ‘cyclic’ coordinate. Conservation of energy constrains r to lie between ‘turning points’, i.e., zerosof E − V eff ( r ) where the radial velocity ˙ r momentarily vanishes. One finds that the orbits are Keplerianellipses for E < along with parabolae and hyperbolae for E ≥ : r ( θ ) = ρ (1 + (cid:15) cos θ ) − [3, 4].Here, ρ = L z /mα is the radius of the circular orbit corresponding to angular momentum L z , (cid:15) theeccentricity and E = − α ρ (1 − (cid:15) ) the energy. Ap×L m ⍺ r /r pA ⊙ θ L aphelionperihelion focus center r Ecliptic Plane
Semi-major axisSemi-minor axis
Figure 1:
Keplerian ellipse in the ecliptic plane of motion showing the constant LRL vector A . The constant angularmomentum L points out of the ecliptic plane. In addition to E and L , the Laplace-Runge-Lenz (LRL) vector A = p × L − mα ˆ r is anotherconstant of motion. It points along the semi-major axis from the CM to the perihelion and its magnitudedetermines the eccentricity of the orbit. Thus, we have conserved quantities: energy and three compo-nents each of L and A . However, a system with three degrees of freedom has a six-dimensional phasespace (space of coordinates and momenta, also called the state space) and if it is to admit continuoustime evolution, it cannot have more than 5 independent conserved quantities. The apparent paradox isresolved once we notice that E , L and A are not all independent; they satisfy two relations : L · A = 0 and E = A − m α m L . (6)Newton used the solution of the two-body problem to understand the orbits of planets and comets. Hethen turned his attention to the motion of the Moon around the Earth. However, lunar motion is signif-icantly affected by the Sun. For instance, A is not conserved and the lunar perigee rotates by ◦ peryear. Thus, he was led to study the Moon-Earth-Sun three-body problem. Wolfgang Pauli (1926) derived the quantum mechanical spectrum of the Hydrogen atom using the relation between E, L and A before the development of the Schr¨odinger equation. Indeed, if we postulate circular Bohr orbits which have zeroeccentricity ( A = 0 ) and quantized angular momentum L = n (cid:126) , then E n = − mα (cid:126) n where α = e / π(cid:15) is theelectromagnetic analogue of Gm m . The three-body problem
We consider the problem of three point masses ( m a with position vectors r a for a = 1 , , ) movingunder their mutual gravitational attraction. This system has 9 degrees of freedom, whose dynamics isdetermined by 9 coupled second order nonlinear ODEs: m a d r a dt = (cid:88) b (cid:54) = a Gm a m b r b − r a | r b − r a | for a = 1 , and . (7)As before, the three components of momentum P = (cid:80) a m a ˙ r a , three components of angular momentum L = (cid:80) a r a × p a and energy E = 12 (cid:88) a =1 m a ˙ r a − (cid:88) a
The 18 phase space variables of the 3-bodyproblem (components of r , r , r , p , p , p ) satisfy 18 first order ordinary differential equations(ODEs) ˙ r a = p a , ˙ p a = −∇ r a V . Lagrange (1772) used the conservation laws to reduce theseODEs to a system of 7 first order ODEs. Conservation of momentum determines 6 phase spacevariables comprising the location R CM and momentum P of the center of mass. Conservation ofangular momentum L = (cid:80) r a × p a and energy E lead to 4 additional constraints. By using one ofthe coordinates as a parameter along the orbit (in place of time), Lagrange reduced the three-bodyproblem to a system of first order nonlinear ODEs. Jacobi vectors (see Fig. 2) generalize the notion of CM and relative coordinates to the 3-body prob-lem [5]. They are defined as J = r − r , J = r − m r + m r m + m and J = m r + m r + m r m + m + m . (9) J is the coordinate of the CM, J the position vector of m relative to m and J that of m relative tothe CM of m and m . A nice feature of Jacobi vectors is that the kinetic energy T = (cid:80) a =1 , , m a ˙ r a and moment of inertia I = (cid:80) a =1 , , m a r a , regarded as quadratic forms, remain diagonal : T = 12 (cid:88) ≤ a ≤ M a ˙ J a and I = (cid:88) ≤ a ≤ M a J a . (10)What is more, just as the potential energy − α/ | r | in the two-body problem is a function only of therelative coordinate r , here the potential energy V may be expressed entirely in terms of J and J : V = − Gm m | J | − Gm m | J − µ J | − Gm m | J + µ J | where µ , = m , m + m . (11) A quadratic form (cid:80) a,b r a Q ab r b is diagonal if Q ab = 0 for a (cid:54) = b . Here, M = m + m is the reduced mass of thefirst pair, M = m + m + m is the reduced mass of m and the ( m , m ) system and M = m + m + m the totalmass. J are cyclic coordinates in the Hamiltonian H = T + V . Inother words, the center of mass motion ( ¨ J = 0 ) decouples from that of J and J .An instantaneous configuration of the three bodies defines a triangle with masses at its vertices. Themoment of inertia about the center of mass I CM = M J + M J determines the size of the triangle.For instance, particles suffer a triple collision when I CM → while I CM → ∞ when one of the bodiesflies off to infinity. CMJ CM m m m r r r J J o Figure 2:
Jacobi vectors J , J and J for the three-body problem. O is the origin of the coordinate system while CM isthe center of mass of particles 1 and 2. The planar three-body problem is the special case where the masses always lie on a fixed plane. Forinstance, this happens when the CM is at rest ( ˙ J = 0 ) and the angular momentum about the CMvanishes ( L CM = M J × ˙ J + M J × ˙ J = 0 ). In 1767, the Swiss scientist Leonhard Euler discoveredsimple periodic solutions to the planar three-body problem where the masses are always collinear, witheach body traversing a Keplerian orbit about their common CM. The line through the masses rotates aboutthe CM with the ratio of separations remaining constant (see Fig. 3a). The Italian/French mathematicianJoseph-Louis Lagrange rediscovered Euler’s solution in 1772 and also found new periodic solutionswhere the masses are always at the vertices of equilateral triangles whose size and angular orientationmay change with time (see Fig. 4). In the limiting case of zero angular momentum, the three bodiesmove toward/away from their CM along straight lines. These implosion/explosion solutions are calledLagrange homotheties.It is convenient to identify the plane of motion with the complex plane C and let the three complexnumbers z a =1 , , ( t ) denote the positions of the three masses at time t . E.g., the real and imaginaryparts of z denote the Cartesian components of the position vector r of the first mass. In Lagrange’ssolutions, z a ( t ) lie at vertices of an equilateral triangle while they are collinear in Euler’s solutions. Inboth cases, the force on each body is always toward the common center of mass and proportional to thedistance from it. For instance, the force on m in a Lagrange solution is F = Gm m r − r | r − r | + Gm m r − r | r − r | = Gm d ( m r + m r + m r − M r ) (12)where d = | r − r | = | r − r | is the side-length of the equilateral triangle and M = m + m + m .Recalling that r CM = ( m r + m r + m r ) /M , we get F = Gm d M ( r CM − r ) ≡ Gm δ r CM − r | r CM − r | (13)6 CMm m (a) Masses traverse Keplerian ellipses with one focus at theCM.
Mm m (b)
Two equal masses m in a circular orbit around a thirdmass M at their CM. Figure 3:
Euler collinear periodic solutions of the three-body problem. The constant ratios of separations are functions of themass ratios alone. m m m m m m CM Figure 4:
Lagrange’s periodic solution with three bodies at vertices of equilateral triangles. The constant ratios of separationsare functions of the mass ratios alone. where δ = M | r CM − r | /d is a function of the masses alone . Thus, the equation of motion for m , m ¨ r = Gm δ r CM − r | r CM − r | , (14)takes the same form as in the two-body Kepler problem (see Eq. 1). The same applies to m and m . Soif z a (0) denote the initial positions, the curves z a ( t ) = z ( t ) z a (0) are solutions of Newton’s equationsfor three bodies provided z ( t ) is a Keplerian orbit for an appropriate two-body problem. In other words,each mass traverses a rescaled Keplerian orbit about the common centre of mass. A similar analysisapplies to the Euler collinear solutions as well: locations of the masses is determined by the requirementthat the force on each one is toward the CM and proportional to the distance from it (see Box 5 on centralconfigurations). Box 5: Central configurations:
Three-body configurations in which the acceleration of each par-ticle points towards the CM and is proportional to its distance from the CM ( a b = ω ( R CM − r b ) for b = 1 , , ) are called ‘central configurations’. A central configuration rotating at angular speed ω about the CM automatically satisfies the equations of motion (7). Euler collinear and Lagrange Indeed, r CM − r = ( m ( r − r ) + m ( r − r )) /M ≡ ( m b + m c ) /M where b and c are two of the sides ofthe equilateral triangle of length d . This leads to | ( r CM − r ) /d | = (cid:112) m + m + m m /M which is a function of massesalone. The restricted three-body problem is a simplified version of the three-body problem where one of themasses m is assumed much smaller than the primaries m and m . Thus, m and m move inKeplerian orbits which are not affected by m . The Sun-Earth-Moon system provides an example wherewe further have m (cid:28) m . In the planar circular restricted three-body problem, the primaries move infixed circular orbits around their common CM with angular speed Ω = ( G ( m + m ) /d ) / given byKepler’s third law and m moves in the same plane as m and m . Here, d is the separation betweenprimaries. This system has degrees of freedom associated to the planar motion of m , and therefore a4-dimensional phase space just like the planar Kepler problem for the reduced mass. However, unlike thelatter which has three conserved quantities (energy, z -component of angular momentum and directionof LRL vector) and is exactly solvable, the planar restricted three-body problem has only one knownconserved quantity, the ‘Jacobi integral’, which is the energy of m in the co-rotating (non-inertial)frame of the primaries: E = (cid:20) m ˙ r + 12 m r ˙ φ (cid:21) − m Ω r − Gm (cid:18) m r + m r (cid:19) ≡ T + V eff . (15)Here, ( r, φ ) are the plane polar coordinates of m in the co-rotating frame of the primaries with originlocated at their center of mass while r and r are the distances of m from m and m (see Fig. 5).The ‘Roche’ effective potential V eff , named after the French astronomer ´Edouard Albert Roche, is a sumof centrifugal and gravitational energies due to m and m . m m m CM (x ,0) (0,0) (x ,0) r r r ϕ Figure 5:
The secondary m in the co-rotating frame of primaries m and m in the restricted three-body problem. Theorigin is located at the center of mass of m and m which coincides with the CM of the system since m (cid:28) m , . A system with n degrees of freedom needs at least n constants of motion to be exactly solvable .For the restricted 3-body problem, Henri Poincar´e (1889) proved the nonexistence of any conservedquantity (other than E ) that is analytic in small mass ratios ( m /m and ( m + m ) /m ) and orbitalelements ( J , M ˙ J , J and M ˙ J ) [6, 7, 8]. This was an extension of a result of Heinrich Bruns whohad proved in 1887 the nonexistence of any new conserved quantity algebraic in Cartesian coordinatesand momenta for the general three-body problem [9]. Thus, roughly speaking, Poincar´e showed that the A Hamiltonian system with n degrees of freedom is exactly solvable in the sense of Liouville if it possesses n independentconserved quantities in involution, i.e., with vanishing pairwise Poisson brackets (see Boxes 6 and 10). §
7, he discovered that itdisplays chaotic behavior.
Euler and Lagrange points (denoted L − ) of the restricted three-body problem are the locationsof a third mass ( m (cid:28) m , m ) in the co-rotating frame of the primaries m and m in the Eulerand Lagrange solutions (see Fig. 6). Their stability would allow an asteroid or satellite to occupy aLagrange point. Euler points L , , are saddle points of the Roche potential while L , are maxima(see Fig. 7). This suggests that they are all unstable. However, V eff does not include the effect of theCoriolis force since it does no work. A more careful analysis shows that the Coriolis force stabilizes L , . It is a bit like a magnetic force which does no work but can stabilize a particle in a Penning trap.Euler points are always unstable while the Lagrange points L , are stable to small perturbations iff ( m + m ) ≥ m m [10]. More generally, in the unrestricted three-body problem, the Lagrangeequilateral solutions are stable iff ( m + m + m ) ≥ m m + m m + m m ) . (16)The above criterion due to Edward Routh (1877) is satisfied if one of the masses dominates the othertwo. For instance, L , for the Sun-Jupiter system are stable and occupied by the Trojan asteroids. m m L L L L L Figure 6:
The positions of Euler ( L , , ) and Lagrange ( L , ) points when m (cid:29) m (cid:29) m . m is in an approximatelycircular orbit around m . L is almost diametrically opposite to m and a bit closer to m than m is. L and L aresymmetrically located on either side of m . L and L are equidistant from m and m and lie on the circular orbit of m . Given the complexity of the restricted three-body problem, Euler (1760) proposed the even simplerproblem of a mass m moving in the gravitational potential of two fixed masses m and m . Initialconditions can be chosen so that m always moves on a fixed plane containing m and m . Thus, wearrive at a one-body problem with two degrees of freedom and energy E = 12 m (cid:0) ˙ x + ˙ y (cid:1) − µ r − µ r . (17)Here, ( x, y ) are the Cartesian coordinates of m , r a the distances of m from m a and µ a = Gm a m for a = 1 , (see Fig. 8). Unlike in the restricted three-body problem, here the rest-frame of the primariesis an inertial frame, so there are no centrifugal or Coriolis forces. This simplification allows the Eulerthree-body problem to be exactly solved. Lagrange points L − are also called libration (literally, balance) points. Stable ‘Halo’ orbits around Euler points have been found numerically. = m = m = .1 m ( .6, 0 ) m (- .4, 0 ) V =- =- =- =- =- =- =- =- =- =- L L L L L G = - - - - Figure 7:
Level curves of the Roche effective potential energy V eff of m in the co-rotating frame of the primaries m and m in the circular restricted three-body problem for G = 1 , m = 15 , m = 10 and m = . . Lagrange points L − are atextrema of V eff . The trajectory of m for a given energy E must lie in the Hill region defined by V eff ( x, y ) ≤ E . E.g., for E = − , the Hill region is the union of two neighborhoods of the primaries and a neighborhood of the point at infinity. Thelobes of the ∞ -shaped level curve passing through L are called Roche’s lobes. The saddle point L is like a mountain passthrough which material could pass between the lobes. Just as the Kepler problem simplifies in plane-polar coordinates ( r, θ ) centered at the CM, the Euler3-body problem simplifies in an elliptical coordinate system ( ξ, η ) . The level curves of ξ and η aremutually orthogonal confocal ellipses and hyperbolae (see Fig. 8) with the two fixed masses at the foci f apart: x = f cosh ξ cos η and y = f sinh ξ sin η. (18)Here, ξ and η are like the radial distance r and angle θ , whose level curves are mutually orthogonalconcentric circles and radial rays. The distances of m from m , are r , = f (cosh ξ ∓ cos η ) .The above confocal ellipses and hyperbolae are Keplerian orbits when a single fixed mass ( m or m ) is present at one of the foci ( ± f, . Remarkably, these Keplerian orbits survive as orbits of theEuler 3-body problem. This is a consequence of Bonnet’s theorem, which states that if a curve is atrajectory in two separate force fields, it remains a trajectory in the presence of both. If v and v are thespeeds of the Keplerian trajectories when only m or m was present, then v = (cid:112) v + v is the speedwhen both are present.Bonnet’s theorem however does not give us all the trajectories of the Euler 3-body problem. Moregenerally, we may integrate the equations of motion by the method of separation of variables in theHamilton-Jacobi equation (see [12] and Boxes 6, 7 & 8). The system possesses two independent con-served quantities: energy and Whittaker’s constant [2, 9] w = L · L + 2 mf ( − µ cos θ + µ cos θ ) = m r r ˙ θ ˙ θ + 2 f m ( − µ cos θ + µ cos θ ) . (19) When the primaries coalesce at the origin ( f → ), Whittaker’s constant reduces to the conserved quantity L of theplanar 2-body problem. = π / ξ = ξ = η = π / η = π / η = π / η = π / η = π / η = η = π m ( f , 0 ) m (- f , 0 ) ξ = ( x,y ) r r η = π / η = π / η = π / - - Figure 8:
Elliptical coordinate system for the Euler 3-body problem. Two masses are at the foci ( ± f, of an ellipticalcoordinate system with f = 2 on the x - y plane. The level curves of ξ and η (confocal ellipses and hyperbolae) are indicated. Here, θ a are the angles between the position vectors r a and the positive x -axis and L , = mr , ˙ θ , ˆ z are the angular momenta about the two force centers (Fig. 8). Since w is conserved, it Poisson commuteswith the Hamiltonian H . Thus, the planar Euler 3-body problem has two degrees of freedom and twoconserved quantities in involution. Consequently, the system is integrable in the sense of Liouville.More generally, in the three-dimensional Euler three-body problem, the mass m can revolve (non-uniformly) about the line joining the force centers ( x -axis) so that its motion is no longer confined to aplane. Nevertheless, the problem is exactly solvable as the equations admit three independent constantsof motion in involution: energy, Whittaker’s constant and the x component of angular momentum [2]. Box 6: Canonical transformations:
We have seen that the Kepler problem is more easily solved inpolar coordinates and momenta ( r, θ, p r , p θ ) than in Cartesian phase space variables ( x, y, p x , p y ) .This change is an example of a canonical transformation (CT). More generally, a CT is a changeof canonical phase space variables ( q , p ) → ( Q ( p , q , t ) , P ( p , q , t )) that preserves the form ofHamilton’s equations. For one degree of freedom, Hamilton’s equations ˙ q = ∂H∂p and ˙ p = − ∂H∂q become ˙ Q = ∂K∂P and ˙ P = − ∂K∂Q where K ( Q, P, t ) is the new Hamiltonian (for a time independentCT, the old and new Hamiltonians are related by substitution: H ( q, p ) = K ( Q ( q, p ) , P ( q, p )) ). Theform of Hamilton’s equations is preserved provided the basic Poisson brackets do not change i.e., { q, p } = 1 , { q, q } = { p, p } = 0 ⇒ { Q, P } = 1 , { Q, Q } = { P, P } = 0 . (20)Here, the Poisson bracket of two functions on phase space f ( q, p ) and g ( q, p ) is defined as { f ( q, p ) , g ( q, p ) } = ∂f∂q ∂g∂p − ∂f∂p ∂g∂q . (21)For one degree of freedom, a CT is simply an area and orientation preserving transformation of the q - p phase plane. Indeed, the condition { Q, P } = 1 simply states that the Jacobian determinant J = det (cid:16) ∂Q∂q , ∂Q∂p | ∂P∂q , ∂P∂p (cid:17) = 1 so that the new area element dQ dP = J dq dp is equal to theold one. A CT can be obtained from a suitable generating function, say of the form S ( q, P, t ) , in11he sense that the equations of transformation are given by partial derivatives of S : p = ∂S∂q , Q = ∂S∂P and K = H + ∂S∂t . (22)For example, S = qP generates the identity transformation ( Q = q and P = p ) while S = − qP generates a rotation of the phase plane by π ( Q = − q and P = − p ). Box 7: Hamilton Jacobi equation:
The Hamilton Jacobi (HJ) equation is an alternative formu-lation of Newtonian dynamics. Let i = 1 , . . . , n label the degrees of freedom of a mechanicalsystem. Cyclic coordinates q i (i.e., those that do not appear in the Hamiltonian H ( q , p , t ) so that ∂H/∂q i = 0 ) help to understand Newtonian trajectories, since their conjugate momenta p i areconserved ( ˙ p i = ∂H∂q i = 0 ). If all coordinates are cyclic, then each of them evolves linearly in time: q i ( t ) = q i (0) + ∂H∂p i t . Now time-evolution is even simpler if ∂H∂p i = 0 for all i as well, i.e., if H is independent of both coordinates and momenta! In the HJ approach, we find a CT from oldphase space variables ( q , p ) to such a coordinate system ( Q , P ) in which the new Hamiltonian K is a constant (which can be taken to vanish by shifting the zero of energy). The HJ equationis a nonlinear, first-order partial differential equation for Hamilton’s principal function S ( q , P , t ) which generates the canonical transformation from ( q , p ) to ( Q , P ) . As explained in Box 6, thismeans p i = ∂S∂q i , Q j = ∂S∂P j and K = H + ∂S∂t . Thus, the HJ equation H (cid:18) q , ∂S∂ q , t (cid:19) + ∂S∂t = 0 (23)is simply the condition for the new Hamiltonian K to vanish. If H is time-independent, we may‘separate’ the time-dependence of S by writing S ( q , P , t ) = W ( q , P ) − Et where the ‘separationconstant’ E may be interpreted as energy. Thus, the time independent HJ-equation for Hamilton’scharacteristic function W is H (cid:18) q , ∂W∂ q (cid:19) = E. (24)E.g., for a particle in a potential V ( q ) , it is the equation m (cid:16) ∂W∂ q (cid:17) + V ( q ) = E . By solving(24) for W , we find the desired canonical transformation to the new conserved coordinates Q andmomenta P . By inverting the relation ( q, p ) (cid:55)→ ( Q, P ) we find ( q i ( t ) , p j ( t )) given their initialvalues. W is said to be a complete integral of the HJ equation if it depends on n constants ofintegration, which may be taken to be the new momenta P , . . . , P n . When this is the case, thesystem is said to be integrable via the HJ equation. However, it is seldom possible to find such acomplete integral. In favorable cases, separation of variables can help to solve the HJ equation (seeBox 8). Box 8: Separation of variables:
In the planar Euler 3-body problem, Hamilton’s characteristicfunction W depends on the two ‘old’ elliptical coordinates ξ and η . The virtue of elliptical coor-dinates is that the time-independent HJ equation can be solved by separating the dependence of W on ξ and η : W ( ξ, η ) = W ( ξ ) + W ( η ) . Writing the energy (17) in elliptical coordinates (18) and12sing p ξ = W (cid:48) ( ξ ) and p η = W (cid:48) ( η ) , the time-independent HJ equation (24) becomes E = W (cid:48) ( ξ ) + W (cid:48) ( η ) − mf ( µ + µ ) cosh ξ − mf ( µ − µ ) cos η mf (cosh ξ − cos η ) . (25)Rearranging, W (cid:48) − Emf cosh ξ − mf ( µ + µ ) cosh ξ = − W (cid:48) − Emf cos η + 2 mf ( µ − µ ) cos η. (26)Since the LHS and RHS are functions only of ξ and η respectively, they must both be equal to a‘separation constant’ α . Thus, the HJ partial differential equation separates into a pair of decoupledODEs for W ( ξ ) and W ( η ) . The latter may be integrated using elliptic functions. Note thatWhittaker’s constant w (19) may be expressed as w = − mf E − α . The importance of the three-body problem lies in part in the developments that arose from attempts tosolve it [6, 7]. These have had an impact all over astronomy, physics and mathematics.Can planets collide, be ejected from the solar system or suffer significant deviations from their Kep-lerian orbits? This is the question of the stability of the solar system. In the th century, Pierre-SimonLaplace and J. L. Lagrange obtained the first significant results on stability. They showed that to first or-der in the ratio of planetary to solar masses ( M p /M S ), there is no unbounded variation in the semi-majoraxes of the orbits, indicating stability of the solar system. Sim´eon Denis Poisson extended this result tosecond order in M p /M S . However, in what came as a surprise, the Romanian Spiru Haretu (1878) over-came significant technical challenges to find secular terms (growing linearly and quadratically in time) inthe semi-major axes at third order! This was an example of a perturbative expansion, where one expandsa physical quantity in powers of a small parameter (here the semi-major axis was expanded in powersof M p /M S (cid:28) ). Haretu’s result however did not prove instability as the effects of his secular termscould cancel out (see Box 9 for a simple example). But it effectively put an end to the hope of provingthe stability/instability of the solar system using such a perturbative approach.The development of Hamilton’s mechanics and its refinement in the hands of Carl Jacobi was stillfresh when the French dynamical astronomer Charles Delaunay (1846) began the first extensive use ofcanonical transformations (see Box 6) in perturbation theory [13]. The scale of his hand calculations isstaggering: he applied a succession of 505 canonical transformations to a th order perturbative treat-ment of the three-dimensional elliptical restricted three-body problem. He arrived at the equation ofmotion for m in Hamiltonian form using pairs of canonically conjugate orbital variables (3 angularmomentum components, the true anomaly, longitude of the ascending node and distance of the ascend-ing node from perigee). He obtained the latitude and longitude of the moon in trigonometric series ofabout terms with secular terms (see Box 9) eliminated. It wasn’t till 1970-71 that Delaunay’s heroiccalculations were checked and extended using computers at the Boeing Scientific Laboratories [13]!The Swede Anders Lindstedt (1883) developed a systematic method to approximate solutions tononlinear ODEs when naive perturbation series fail due to secular terms (see Box 9). The technique wasfurther developed by Poincar´e. Lindstedt assumed the series to be generally convergent, but Poincar´esoon showed that they are divergent in most cases. Remarkably, nearly 70 years later, Kolmogorov,Arnold and Moser showed that in many of the cases where Poincar´e’s arguments were inconclusive, the13eries are in fact convergent, leading to the celebrated KAM theory of integrable systems subject to smallperturbations (see Box 10). Box 9: Poincar´e-Lindstedt method:
The Poincar´e-Lindstedt method is an approach to findingseries solutions to a system such as the anharmonic oscillator ¨ x + x + gx = 0 , which for small g ,is a perturbation of the harmonic oscillator m ¨ x + kx = 0 with mass m = 1 and spring constant k = 1 . The latter admits the periodic solution x ( t ) = cos t with initial conditions x (0) = 1 , ˙ x (0) = 0 . For a small perturbation < g (cid:28) , expanding x ( t ) = x ( t ) + gx ( t ) + · · · in powersof g leads to a linearized equation for x ( t )¨ x + x + cos t = 0 . (27)However, the perturbative solution x ( t ) = x + gx + O ( g ) = cos t + g (cid:20)
132 (cos 3 t − cos t ) − t sin t (cid:21) + O ( g ) (28)is unbounded due to the linearly growing secular term ( − / t sin t . This is unacceptable as theenergy E = ˙ x + x + gx must be conserved and the particle must oscillate between turningpoints of the potential V = x + g x . The Poincar´e-Lindstedt method avoids this problem bylooking for a series solution of the form x ( t ) = x ( τ ) + g ˜ x ( τ ) + · · · (29)where τ = ωt with ω = 1 + gω + · · · . The constants ω , ω , · · · are chosen to ensure that thecoefficients of the secular terms at order g, g , · · · vanish. In the case at hand we have x ( t ) = cos( t + gω t )+ g ˜ x ( t )+ O ( g ) = cos t + g ˜˜ x ( t )+ O ( g ) where ˜˜ x ( t ) = ˜ x ( t ) − ω t sin t. (30) ˜˜ x satisfies the same equation (27) as x did, leading to ˜ x ( t ) = 132 (cos 3 t − cos t ) + (cid:18) ω − (cid:19) t sin t. (31)The choice ω = 3 / ensures cancellation of the secular term at order g , leading to the approximatebounded solution x ( t ) = cos (cid:18) t + 38 gt (cid:19) + g
32 (cos 3 t − cos t ) + O (cid:0) g (cid:1) . (32) Box 10: Action-angle variables and invariant tori:
Time evolution is particularly simple if allthe generalized coordinates θ j are cyclic so that their conjugate momenta I j are conserved: ˙ I j = − ∂H∂θ j = 0 . A Hamiltonian system with n degrees of freedom is integrable in the sense of Liouvilleif it admits n canonically conjugate ( { θ j , I k } = δ jk a ) pairs of phase space variables ( θ j , I j ) withall the θ j cyclic, so that its Hamiltonian depends only on the momenta, H = H ( I ) . Then the‘angle’ variables θ j evolve linearly in time ( θ j ( t ) = θ j (0) + ω j t ) while the momentum or ‘action’variables I j are conserved. Here, ω j = ˙ θ j = ∂H∂I j are n constant frequencies. Typically the anglevariables are periodic, so that the θ j parametrize circles. The common level sets of the action14ariables I j = c j are therefore a family of tori that foliate the phase space. Recall that a torus is aCartesian product of circles. For instance, for one degree of freedom, θ labels points on a circle S while for 2 degrees of freedom, θ and θ label points on a 2-torus S × S which looks like avada or doughnut. Trajectories remain on a fixed torus determined by the initial conditions. Under asufficiently small and smooth perturbation H ( I )+ gH (cid:48) ( I , (cid:126)θ ) , Andrei Kolmogorov, Vladimir Arnoldand J¨urgen Moser showed that some of these ‘invariant’ tori survive provided the frequencies ω i aresufficiently ‘non-resonant’ or ‘incommensurate’ (i.e., their integral linear combinations do not get‘too small’). a The Kronecker symbol δ jk is equal to one for j = k and zero otherwise George William Hill was motivated by discrepancies in lunar perigee calculations. His celebrated pa-per on this topic was published in 1877 while working with Simon Newcomb at the American Ephemerisand Nautical Almanac . He found a new family of periodic orbits in the circular restricted (Sun-Earth-Moon) 3-body problem by using a frame rotating with the Sun’s angular velocity instead of that of theMoon. The solar perturbation to lunar motion around the Earth results in differential equations with pe-riodic coefficients. He used Fourier series to convert these ODEs to an infinite system of linear algebraicequations and developed a theory of infinite determinants to solve them and obtain a rapidly convergingseries solution for lunar motion. He also discovered new ‘tight binary’ solutions to the 3-body problemwhere two nearby masses are in nearly circular orbits around their center of mass CM , while CM and the far away third mass in turn orbit each other in nearly circular trajectories.The French mathematician/physicist/engineer Henri Poincar´e began by developing a qualitative the-ory of differential equations from a global geometric viewpoint of the dynamics on phase space. Thisincluded a classification of the types of equilibria (zeros of vector fields) on the phase plane (nodes, sad-dles, foci and centers, see Fig. 9). His 1890 memoir on the three-body problem was the prize-winningentry in King Oscar II’s th birthday competition (for a detailed account see [8]). He proved the di-vergence of series solutions for the 3-body problem developed by Delaunay, Hugo Gyld´en and Lindstedt(in many cases) and covergence of Hill’s infinite determinants. To investigate the stability of 3-body mo-tions, Poincar´e defined his ‘surfaces of section’ and a discrete-time dynamics via the ‘return map’ (seeFig. 10). A Poincar´e surface S is a two-dimensional surface in phase space transversal to trajectories.The first return map takes a point q on S to q , which is the next intersection of the trajectory through q with S . Given a saddle point p on a surface S , he defined its stable and unstable spaces W s and W u as points on S that tend to p upon repeated forward or backward applications of the return map (seeFig. 11). He initially assumed that W s and W u on a surface could not intersect and used this to arguethat the solar system is stable. This assumption turned out to be false, as he discovered with the helpof Lars Phragm´en. In fact, W s and W u can intersect transversally on a surface at a homoclinic point if the state space of the underlying continuous dynamics is at least three-dimensional. What is more,he showed that if there is one homoclinic point, then there must be infinitely many accumulating at p .Moreover, W s and W u fold and intersect in a very complicated ‘homoclinic tangle’ in the vicinity of p . This was the first example of what we now call chaos. Chaos is usually manifested via an extremesensitivity to initial conditions (exponentially diverging trajectories with nearby initial conditions).When two gravitating point masses collide, their relative speed diverges and solutions to the equationsof motion become singular at the collision time t c . More generally, a singularity occurs when either a Simon Newcomb’s project of revising all the orbital data in the solar system established the missing (cid:48)(cid:48) in the (cid:48)(cid:48) centennial precession of Mercury’s perihelion. This played an important role in validating Einstein’s general theory of relativity. Homoclinic refers to the property of being ‘inclined’ both forward and backward in time to the same point. a) center (b) (stable) node (c) (unstable) focus (d) saddle Figure 9:
Poincar´e’s classification of zeros of a vector field (equilibrium or fixed points) on a plane. (a) Center is alwaysstable with oscillatory motion nearby, (b,c) nodes and foci (or spirals) can be stable or unstable and (d) saddles are unstableexcept in one direction. q q S Figure 10:
A Poincare surface S transversal to a trajectory is shown. The trajectory through q on S intersects S again at q . The map taking q to q is called Poincar´e’s first return map. position or velocity diverges in finite time. The Frenchman Paul Painlev´e (1895) showed that binary andtriple collisions are the only possible singularities in the three-body problem. However, he conjecturedthat non-collisional singularities (e.g. where the separation between a pair of bodies goes to infinityin finite time) are possible for four or more bodies. It took nearly a century for this conjecture to beproven, culminating in the work of Donald Saari and Zhihong Xia (1992) and Joseph Gerver (1991)who found explicit examples of non-collisional singularities in the -body and n -body problems for n sufficiently large [14]. In Xia’s example, a particle oscillates with ever growing frequency and amplitudebetween two pairs of tight binaries. The separation between the binaries diverges in finite time, as doesthe velocity of the oscillating particle.The Italian mathematician Tulio Levi-Civita (1901) attempted to avoid singularities and thereby ‘reg-ularize’ collisions in the three-body problem by a change of variables in the differential equations. Forexample, the ODE for the one-dimensional Kepler problem ¨ x = − k/x is singular at the collisionpoint x = 0 . This singularity can be regularized by introducing a new coordinate x = u and areparametrized time ds = dt/u , which satisfy the nonsingular oscillator equation u (cid:48)(cid:48) ( s ) = Eu/ with conserved energy E = (2 ˙ u − k ) /u . Such regularizations could shed light on near-collisionaltrajectories (‘near misses’) provided the differential equations remain physically valid .The Finnish mathematician Karl Sundman (1912) began by showing that binary collisional singular-ities in the 3-body problem could be regularized by a repararmetrization of time, s = | t − t | / where t is the the binary collision time [15]. He used this to find a convergent series representation (in powersof s ) of the general solution of the 3-body problem in the absence of triple collisions . The possibilityof such a convergent series had been anticipated by Karl Weierstrass in proposing the 3-body problem Solutions which could be smoothly extended beyond collision time (e.g., the bodies elastically collide) were called regu-larizable. Those that could not were said to have an essential or transcendent singularity at the collision. Note that the point particle approximation to the equations for celestial bodies of non-zero size breaks down due to tidaleffects when the bodies get very close Sundman showed that for non-zero angular momentum, there are no triple collisions in the three-body problem. -1 h h W u W s p Figure 11:
The saddle point p and its stable and unstable spaces W s and W u are shown on a Poincar´e surface through p . The points at which W s and W u intersect are called homoclinic points, e.g., h , h and h − . Points on W s (or W u )remain on W s (or W u ) under forward and backward iterations of the return map. Thus, the forward and backward images of ahomoclinic point under the return map are also homoclinic points. In the figure h is a homoclinic point whose image is h onthe segment [ h , p ] of W s . Thus, W u must fold back to intersect W s at h . Similarly, if h − is the backward image of h on W u , then W s must fold back to intersect W u at h − . Further iterations produce an infinite number of homoclinic pointsaccumulating at p . The first example of a homoclinic tangle was discovered by Poincar´e in the restricted 3-body problem andis a signature of its chaotic nature. for King Oscar’s 60th birthday competition. However, Sundman’s series converges exceptionally slowlyand has not been of much practical or qualitative use.The advent of computers in the th century allowed numerical investigations into the 3-body (andmore generally the n -body) problem. Such numerical simulations have made possible the accurateplacement of satellites in near-Earth orbits as well as our missions to the Moon, Mars and the outerplanets. They have also facilitated theoretical explorations of the three-body problem including chaoticbehavior, the possibility for ejection of one body at high velocity (seen in hypervelocity stars [16]) andquite remarkably, the discovery of new periodic solutions. For instance, in 1993, Chris Moore discoveredthe zero angular momentum figure-8 ‘choreography’ solution. It is a stable periodic solution with bodiesof equal masses chasing each other on an ∞ -shaped trajectory while separated equally in time (seeFig. 12). Alain Chenciner and Richard Montgomery [17] proved its existence using an elegant geometricreformulation of Newtonian dynamics that relies on the variational principle of Euler and Maupertuis. m m m Figure 12:
Equal-mass zero-angular momentum figure-8 choreography solution to the 3-body problem. A choreography is aperiodic solution where all masses traverse the same orbit separated equally in time.
Fermat’s principle in optics states that light rays extremize the optical path length (cid:82) n ( r ( τ )) dτ where n ( r ) is the (position dependent) refractive index and τ a parameter along the path . The variationalprinciple of Euler and Maupertuis (1744) is a mechanical analogue of Fermat’s principle [18]. It statesthat the curve that extremizes the abbreviated action (cid:82) q q p · d q holding energy E and the end-points The optical path length (cid:82) n ( r ) dτ is proportional to (cid:82) dτ /λ , which is the geometric length in units of the local wavelength λ ( r ) = c/n ( r ) ν . Here, c is the speed of light in vacuum and ν the constant frequency. and q fixed has the same shape as the Newtonian trajectory. By contrast, Hamilton’s principle ofextremal action (1835) states that a trajectory going from q at time t to q at time t is a curve thatextremizes the action .It is well-known that the trajectory of a free particle (i.e., subject to no forces) moving on a plane is astraight line. Similarly, trajectories of a free particle moving on the surface of a sphere are great circles.More generally, trajectories of a free particle moving on a curved space (Riemannian manifold M ) aregeodesics (curves that extremize length). Precisely, for a mechanical system with configuration space M and Lagrangian L = m ij ( q ) ˙ q i ˙ q j , Lagrange’s equations dp i dt = ∂L∂q i are equivalent to the geodesicequations with respect to the ‘kinetic metric’ m ij on M : m ij ¨ q j ( t ) = −
12 ( m ji,k + m ki,j − m jk,i ) ˙ q j ( t ) ˙ q k ( t ) . (33)Here, m ij,k = ∂m ij /∂q k and p i = ∂L∂ ˙ q i = m ij ˙ q j is the momentum conjugate to coordinate q i . Forinstance, the kinetic metric ( m rr = m , m θθ = mr , m rθ = m θr = 0 ) for a free particle moving on aplane may be read off from the Lagrangian L = m ( ˙ r + r ˙ θ ) in polar coordinates, and the geodesicequations shown to reduce to Lagrange’s equations of motion ¨ r = r ˙ θ and d ( mr ˙ θ ) /dt = 0 .Remarkably, the correspondence between trajectories and geodesics continues to hold even in thepresence of conservative forces derived from a potential V . Indeed, trajectories of the Lagrangian L = T − V = m ij ( q ) ˙ q i ˙ q j − V ( q ) are reparametrized geodesics of the Jacobi-Maupertuis (JM) metric g ij = ( E − V ( q )) m ij ( q ) on M where E = T + V is the energy. This geometric formulation of theEuler-Maupertuis principle (due to Jacobi) follows from the observation that the square of the metric lineelement ds = g ij dq i dq j = ( E − V ) m ij dq i dq j = 12 m kl dq k dt dq l dt m ij dq i dq j = 12 (cid:0) m ij ˙ q i dq j (cid:1) = 12 ( p · d q ) , (34)so that the extremization of (cid:82) p · d q is equivalent to the extremization of arc length (cid:82) ds . Loosely, thepotential V ( q ) on the configuration space plays the role of an inhomogeneous refractive index. Thoughtrajectories and geodesics are the same curves, the Newtonian time t along trajectories is in generaldifferent from the arc-length parameter s along geodesics. They are related by dsdt = √ E − V ) [19].This geometric reformulation of classical dynamics allows us to assign a local curvature to pointson the configuration space. For instance, the Gaussian curvature K of a surface at a point (see Box11) measures how nearby geodesics behave (see Fig. 13), they oscillate if K > (as on a sphere),diverge exponentially if K < (as on a hyperboloid) and linearly separate if K = 0 (as on a plane).Thus, the curvature of the Jacobi-Maupertuis metric defined above furnishes information on the stabilityof trajectories. Negativity of curvature leads to sensitive dependence on initial conditions and can be asource of chaos. The action is the integral of the Lagrangian S = (cid:82) t t L ( q , ˙ q ) dt . Typically, L = T − V is the difference between kineticand potential energies. A metric m ij on an n -dimensional configuration space M is an n × n matrix at each point q ∈ M that determinesthe square of the distance ( ds = (cid:80) ni,j =1 m ij dq i dq j ) from q to a nearby point q + d q . We often suppress the summationsymbol and follow the convention that repeated indices are summed from to n . The shapes of trajectories and geodesics coincide but the Newtonian time along trajectories is not the same as the arc-lengthparameter along geodesics. a) Nearby geodesics on a plane ( K =0 ) separate linearly. (b) Distance between neighboringgeodesics on a sphere (
K > )oscillates. (c) Geodesics on a hyperbolic surface(
K < ) deviate exponentially Figure 13:
Local behavior of nearby geodesics on a surface depends on the sign of its Gaussian curvature K . Box 11: Gaussian curvature:
Given a point p on a surface S embedded in three dimensions,a normal plane through p is one that is orthogonal to the tangent plane at p . Each normal planeintersects S along a curve whose best quadratic approximation at p is called its osculating circle.The principal radii of curvature R , at p are the maximum and minimum radii of osculating circlesthrough p . The Gaussian curvature K ( p ) is defined as /R R and is taken positive if the centersof the corresponding osculating circles lie on the same side of S and negative otherwise.In the planar Kepler problem, the Hamiltonian (5) in the CM frame is H = p x + p y m − αr where α = GM m > and r = x + y . (35)The corresponding JM metric line element in polar coordinates is ds = m (cid:0) E + αr (cid:1) (cid:0) dr + r dθ (cid:1) . ItsGaussian curvature K = − Eα/ m ( α + Er ) has a sign opposite to that of energy everywhere. Thisreflects the divergence of nearby hyperbolic orbits and oscillation of nearby elliptical orbits. Despitenegativity of curvature and the consequent sensitivity to initial conditions, hyperbolic orbits in the Keplerproblem are not chaotic: particles simply fly off to infinity and trajectories are quite regular. On theother hand, negativity of curvature without any scope for escape can lead to chaos. This happens withgeodesic motion on a compact Riemann surface with constant negative curvature: most trajectories arevery irregular. We now sketch how the above geometrical framework may be usefully applied to the three-body problem.The configuration space of the planar 3-body problem is the space of triangles on the plane with massesat the vertices. It may be identified with six-dimensional Euclidean space ( R ) with the three planarJacobi vectors J , , (see (9) and Fig. 2) furnishing coordinates on it. A simultaneous translation of theposition vectors of all three bodies r , , (cid:55)→ r , , + r is a symmetry of the Hamiltonian H = T + V of Eqs. (10,11) and of the Jacobi-Maupertuis metric ds = ( E − V ( J , J )) (cid:88) a =1 M a | d J a | . (36) A compact Riemann surface is a closed, oriented and bounded surface such as a sphere, a torus or the surface of a pretzel.The genus of such a surface is the number of handles: zero for a sphere, one for a torus and two or more for higher handle-bodies. Riemann surfaces with genus two or more admit metrics with constant negative curvature. -m collisionEuler pointsLagrange points SyzygyLinesm -m collisionm -m collision (a) The negatively curved ‘pair of pants’ metric on the shapesphere S . (η=π/2)1-3 collision(π/3,π/2)2-3 collision(π/3,0) ηξ E (η=π/6,ξ =π/2)E (π/6,0)syzygy great circleLagrange point L (π/4,3π/4)Lagrange point L (π/4,π/4) (b) Locations of Lagrange, Euler and collision points on ageometrically unfaithful depiction of the shape sphere S .The negative curvature of S is indicated in Fig. 14a. Syzy-gies are instantaneous configurations where the three bodiesare collinear (eclipses). Figure 14: ‘Pair of pants’ metric on shape sphere and Lagrange, Euler and collision points.
This is encoded in the cyclicity of J . Quotienting by translations allows us to define a center of massconfiguration space R (the space of centered triangles on the plane with masses at the vertices) withits quotient JM metric. Similarly, rotations J a → (cid:18) cos θ − sin θ sin θ cos θ (cid:19) J a for a = 1 , , are a symmetryof the metric, corresponding to rigid rotations of a triangle about a vertical axis through the CM. Thequotient of R by such rotations is the shape space R , which is the space of congruence classes ofcentered oriented triangles on the plane. Translations and rotations are symmetries of any central inter-particle potential, so the dynamics of the three-body problem in any such potential admits a consistentreduction to geodesic dynamics on the shape space R . Interestingly, for an inverse-square potential (asopposed to the Newtonian ‘ /r ’ potential) V = − (cid:88) a , I → ∞ as t → ∞ so that bodies fly apart asymptotically. Similarly, when E < they suffer a triple collision. When E = 0 , the sign of ˙ I (0) controls asymptotic behaviorleaving open the special case when E = 0 and ˙ I (0) = 0 . By contrast, for the Newtonian potential,the Hamiltonian transforms as H ( λ − / r , λ / p ) = λ / H ( r , p ) leading to the Lagrange-Jacobiidentity ¨ I = 4 E − V . This is however not adequate to determine the long-time behavior of I when E < . References [1] Laskar, J.,
Is the Solar System stable?
Progress in Mathematical Physics, , 239-270 (2013).[2] Gutzwiller, M. C., Chaos in Classical and Quantum mechanics , Springer-Verlag, New York (1990).[3] Goldstein, H., Poole, C. P., and Safko, J. L.,
Classical Mechanics , 3rd Ed., Pearson Education(2011).[4] Hand, L. N. and Finch, J. D.,
Analytical Mechanics , Cambridge Univ. Press (1998).[5] Rajeev, S. G.,
Advanced Mechanics: From Euler’s Determinism to Arnold’s Chaos , Oxford Uni-versity Press, Oxford (2013).[6] Diacu F. and Holmes P.,
Celestial Encounters: The Origins of Chaos and Stability , Princeton Uni-versity Press, New Jersey (1996). Scalar curvature is an average of the Gaussian curvatures in the various tangent planes through a point
The three-body problem , Reports on Progress in Physics, , 6,065901 (2014), arXiv:1508.02312.[8] Barrow-Green, J., Poincar´e and the Three Body Problem , Amer. Math. Soc., Providence, RhodeIsland (1997).[9] Whittaker, E. T.,
A treatise on the analytical dynamics of particles & rigid bodies , 2nd Ed., Cam-bridge University Press, Cambridge (1917), Chapt. XIV and page 283.[10] Symon, K. R.,
Mechanics , 3rd Ed., Addison Wesley, Philippines (1971).[11] Bodenmann, S.,
The 18th-century battle over lunar motion , Physics Today, (1), 27 (2010).[12] Mukunda, N., Sir William Rowan Hamilton , Resonance, (6), 493 (2016).[13] Gutzwiller, M. C., Moon-Earth-Sun: The oldest three-body problem , Reviews of Modern Physics, , 589 (1998).[14] Saari, D. G. and Xia, Z., Off to infinity in finite time , Notices of the AMS, , 538 (1993).[15] Siegel, C.L. and Moser, J.K., Lectures on Celestial Mechanics , Springer-Verlag, Berlin (1971),page 31.[16] Brown, W. R.,
Hypervelocity Stars in the Milky Way , Physics Today, (6), 52 (2016).[17] Montgomery, R., A new solution to the three-body problem , Notices of the AMS, (5), 471 (2001).[18] Lanczos, C., The variational principles of mechanics , 4th Ed., Dover, New York (1970), page 139.[19] Krishnaswami, G. S. and Senapati, H.,
Curvature and geodesic instabilities in a geometrical ap-proach to the planar three-body problem , J. Math. Phys.,57