4
$\begingroup$

I'm trying to understand a very basic property of Hamiltonian dynamics. I don't have a physics background but I do know some mathematics. I want to understand why negating the momentum is equivalent to reversing time in Hamiltonian dynamics.

Suppose I have a Hamiltonian $H(q, p)$ which satisfies $H(q, p)=H(q, -p)$. The Hamiltonian equations of motion are $$ \dot{p} = -\nabla_q H(q, p) ~~~~~~~~ \dot{q} = \nabla_p H(q, p) $$ That if $(q(t), p(t))$ satisfy Hamilton's equations then so does $(q(-t), -p(-t))$ seems to be an oft-quoted fact in many works on classical mechanics. But how do I convince myself that it is true?

$\endgroup$

1 Answer 1

3
$\begingroup$

If we set $p\rightarrow -p$ then the Hamilton equations look like $$-\dot p=-\nabla_qH(q,p)~~~~~~~~~~~\dot q=\nabla_{-p}H(q,p)=-\nabla_pH(q,p)$$ Now we use the fact that $-\dot q=-\partial q/\partial t=\partial q/\partial (-t) $ to get the modified equations that look like $$\frac{\partial p}{\partial(-t)}= -\nabla_qH(q,p)~~~~~~~~~~~\frac{\partial q}{\partial(-t)}=\nabla_pH(q,p) $$ These are the same equations albeit $t\rightarrow-t$. Thus we see that Hamilton’s equations are symmetric under time reversal and also that this can be brought about by setting $p\rightarrow-p$.

$\endgroup$
2
  • $\begingroup$ When you set $p$ to $-p$ why doesn't the equation for $\dot{q}$ transform according to $\dot{q} = \nabla_{-p} H(q, -p)$? $\endgroup$ Commented Apr 11, 2020 at 17:53
  • $\begingroup$ Because all naturally occurring Hamiltonians have a $p^2$ dependence. $\endgroup$ Commented Apr 12, 2020 at 5:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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