I'm studying calculus of variation in this semester. I have difficulties in understanding the conditions for a curve to be an extremal.

Specifically, when we are deriving the general variation of a functional of the form $\int_{a}^{b}F(x,y,y')\mathrm{d}x$, we define $p$ to be $\frac{\partial F}{\partial y'}$ and $H$ to be $y'\frac{\partial F}{\partial y'}-F$. If we don't put any restrictions on the two endpoints, we must have a) Euler-Lagrange equation must be satisfied; b) $p\delta_y|_a^b - H\delta_x|_a^b=0$.

Our professor mentioned that $p$ and $H$ represent momentum and hamiltonian respectively. But I can't understand the relationship between these two quantities and the variational problem. Because I've never seen any variational problem which involves momentum $p$ and hamiltonian $H$. An example from physics which involves momentum and hamiltonian would be very helpful.

--------The followings are edited on Apr 4, 2020---------

Moreover, I would like to know the intuition of the quantity $\delta J = p\delta_y|_{a}^{b}-H\delta_x|_{a}^{b}$ where $J[y] = \int_{a}^{b}F(x,y,y')$. This corresponds to the situation where the endpoints are not fixed and the curve $y$ is taken to be the extremal (satisfies the Euler-Lagrange equations). How could this quantity be interpreted as the variation when the endpoints are perturbed? Need a concrete example to illustrate this.

  • $\begingroup$ What's the definition of $\delta_x$ and $\delta_y$? $\endgroup$ – Qmechanic Feb 15 at 12:56
  • $\begingroup$ They represent the perturbations of positions of endpoints. $\endgroup$ – Chris Jing Feb 15 at 23:59
  • $\begingroup$ Consider to spell out more clearly the logic. Condition (b) seems to be part of Noether's theorem but the post (v2) has no mentioning of a symmetry per se. $\endgroup$ – Qmechanic Feb 16 at 8:20

Let's discuss a generalization of @Botond's example,$$L=\frac12m\dot{x}^2-V(x)\implies p=m\dot{x}\implies H=\frac12m\dot{x}^2+V(x)$$(this even works in multiple dimensions). Then $L$ is kinetic energy minus potential energy, while $H$ is kinetic plus potential. In other words, $H$ is the total energy. Indeed, $H$ is conserved much more generally, namely whenever $\partial_tL=0$, because$$\frac{dH}{dt}=\frac{\partial H}{\partial t}=-\frac{\partial L}{\partial t}.$$So $H$ is a popular definition of energy.

| cite | improve this answer | |
  • 1
    $\begingroup$ Thank you! Got it now. $\endgroup$ – Chris Jing Apr 5 at 2:39

In quantum mechanics, Schrödinger's equation uses the Hamiltonian as an operator, so you can see there the Hamiltonian almost every time.

But let me show you a simpler example: Imagine a point particle at the end of an ideal spring (simple harmonic oscillator). The mass of the particle is $m$, and the spring constant is $k$. We know that the energy of the system will be $E=\frac{1}{2}mv^2+\frac{1}{2}kx^2$ with momentum $p=mv$, but we can get it directly from the Lagrange function: $$L=\frac{1}{2}mv^2-\frac{1}{2}kx^2$$ The generalised momentum (which happens to equal to the momentum this time) is $$p=\frac{\partial L}{\partial v}=mv$$ So the Hamiltonian is $$H=pv-L=pv-\frac{1}{2}mv^2+\frac{1}{2}kx^2=p\frac{p}{m}-\frac{1}{2}m\frac{p^2}{m^2}+\frac{1}{2}kx^2=\frac{p^2}{2m}+\frac{1}{2}m\omega^2x^2$$ Where we defined $\omega$ as $\omega=\sqrt{\frac{k}{m}}$ (It's nit neccessary, but people usually do this).

| cite | improve this answer | |
  • $\begingroup$ Thank you very much! This was a very good example that I saw in the subsequent classes. $\endgroup$ – Chris Jing Apr 5 at 2:39

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.