Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let be given a mechanical system with configuration manifold $M,$ potential energy $V$ and kinetic energy $K$ corresponding to a riemannian metric on $M.$ Its dynamics is determined by the Euler-Lagrange vector field on $TM$ corresponding to $L=K-V.$

For a given value $e$ of the total energy $K+V,$ let us denote by $n(e)$ the number of closed trajectories of motion with energy, (here I am identifying motions differing just by a time translation).

Question.1 Under what hypothesis on the whole mechanical system, or only on the value $e$, is it possible to give a positive lower bound on $n(e)$?

I was motivated even for having been puzzled, probably for lacking preparation, by a statement in ยง4.2.1 in Arnol'd, Kozlov, Neishtadt Mathematical Aspects of Classical and Celestial Mechanics, (3rd Edn).
There I found that, invoking Hadamard(1898) (in any homotopy free class of a not simply connected riemannian manifold there exists a closed geodesic) and Maupertuis' principle, when $M$ is not simply connected and $\sup_M V<e,$ it is possile to extimate from below $n(e).$

Question.2 What kind of lower bound they are referring to? Probably it is elementary, and I am missing something, but what?

Excuse me if the question is not well-posed, any suggestion in order to improve the terms of the problem are welcome, just as the answers of course.

share|cite|improve this question
Haven't worked out the details, but most likely the idea is to directly generalise Hadamard's theorem (which is one about calculus of variations) into minimising the "length function" $\int_\gamma ds \sqrt{e - V\circ \gamma}$. This maybe even cast as a Finsler geometry (I'm not sure) in which case Hadamard's theorem can be applied directl. The constraint $\sup V < e$ is just so that the for constant energy curves, you cannot visit the points of potential energy higher than the total energy, and thus the sublevel set of $V < e$ may not be the whole manifold and have different topology. – Willie Wong Feb 10 '12 at 8:02
@Willie Wong, My confusion is: ok, in any homotopy free class there exists a geodesic (for the Jacobi metric $(e-V)g,$ and so it is a closed trajectory with energy $e$ for the mechanical system), but I cannot conclude $n(e)\geq|\pi_\text{free}(M)|$. For example, if $\gamma$ is a closed geodesic, then the last term distinguish $[\gamma]^n$ for different $n,$( or not?) – Giuseppe Feb 10 '12 at 10:34
Ah, no, you cannot get $|\pi_{\texrm{free}}(M)|$. A very rough bound you can get is a lower bound by the number of generators of the free homotopy group. Simples example, let $V = 0$ and $M = \mathbb{S}^1$. The only closed geodesic for a fixed energy up to reparametrisation is the circle $\gamma$. But the fundamental group is $\mathbb{Z}$. But this can vastly underestimate: For example, think again about $\mathbb{T}^2$. There's probably some facts about group theory which may help, but I am not familiar enough with them to say. Sorry. – Willie Wong Feb 10 '12 at 13:11
Dear Willie Wong, thank you very much. – Giuseppe Feb 10 '12 at 13:19
You may be interested in this paper, the references therein, and so forth. Leave this question here for a while. If you don't get better answers, you should consider asking on MathOverflow. – Willie Wong Feb 10 '12 at 13:47

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.