How can differential equations describing a physical object's true spatial paths (orbits) be derived from the time-independent Jacobi-Maupertuis principle of least action? According to this, it is possible.

The time-independent version of Maupertuis' principle yields (Lanczos 1970, Landau and Lifshitz 1969) corresponding differential equations for the true spatial paths (orbits).

I guess I am looking for equations analogous to the Euler-Lagrange conditions, but for true spatial paths (orbits) instead of true trajectories as a function of time.

This PSE post is relevant, but is concerned with the derivation of equations of motion, whereas here I am concerned with the derivation of true spatial paths (orbits).

  • 1
    $\begingroup$ Comment to the question (v1): It seems the question formulation is mixing the following non-standard terminology used in the linked Scholarpedia article: 1. a true path = a parametrized on-shell curve, i.e. a function of time; and 2. a true orbit = an un-parametrized on-shell curve, i.e. not a function of time. $\endgroup$ – Qmechanic Dec 15 '15 at 13:59
  • $\begingroup$ @Qmechanic I have tried to address your comment with an edit. $\endgroup$ – ben Dec 15 '15 at 16:42

Comments to the question (v4):

  1. Recall first of all, that in the principle of stationary action, the initial and final times, $t_i$ and $t_f$, are fixed; while in the abbreviated action principle, the initial and final times, $t_i$ and $t_f$, are free, but the initial and final positions, $q^j_i$ and $q^j_f$, together with the energy $E$, are fixed.

  2. There exist reparametrization-invariant formalisms. However, let us here just outline an elementary pedestrian approach. This consists in picking one of the position coordinates (which possibly after a relabelling we call) $q^1\equiv \lambda$ to play the role of the curve parameter $\lambda$. The other position coordinates $q^2, \ldots, q^n$ are now viewed as functions of $q^1$.

  3. In other words, we view the curve as a graph. The bad news is that this might only work locally, and we may have to patch solutions together globally.

  4. The good news is that $q^1\equiv \lambda$ is naturally fixed at the initial and final position, $q^1_i$ and $q^1_f$.

  5. Now let there be given one of the Maupertuis-like action functional, say $A$. After some straightforward manipulations, $A$ becomes of the form $$A[q^2, \ldots, q^n]~=~ \int_{q^1_i}^{q^1_f} d\lambda~ f\left(E,\lambda, q, \frac{dq}{d\lambda}\right) .$$ It should be straightforward to see which equations describe the stationary points of $A$!

| cite | improve this answer | |
  • $\begingroup$ Ok, but it sort of seems like you have just trivially relabeled the time parameter as a space parameter. $\endgroup$ – ben Dec 29 '15 at 15:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.