0
$\begingroup$

Suppose that we are given an extremal problem:

$$\begin{cases}J(x)=\int_{t_0}^{t_1} L(t,x,x') dt \rightarrow inf\\ x(t_0)=x_0 \\ x(t_1)=x_1 \end{cases}$$

And we are asked to find the corresponding Hamilton-Jacobi equation.

I know that the Hamilton-Jacobi equation is of the form:$H + \frac{\partial S}{\partial t}= 0 $

My questions are:

  1. In order to find the Hamiltonian in the past, I have used a process to transform the Euler-Lagrange function into a Hamiltonian system. In the case of this problem, would it be required for me to transform the Jacobi equation into a Hamiltonian system?:

$$-\frac{d}{dt}(Ah' + Ch) +Ch' + Bh = 0\\A=L_{x'x'}\\B=L_{x'x}\\C=L_{xx}\\\vec{h}$$

  1. I am also not exactly sure how to go about formulating the Hamilton principal function either. If someone could give me a better idea of the process on how to approach this problem that would be great. Thanks.
$\endgroup$
0
$\begingroup$

To make formulas a little bit cleaner, at some points I'll use the notation $f_t$ to represent the value $f(t)$ of a function $f$ at $t$. It's convenient to consider the $J$ as both a function of $x$ and the upper limit of the integral $t$: \begin{equation} J(x,t)=\int_{t_0}^tL(x_\tau,x'_\tau,\tau)d\tau. \end{equation} Define the conjugate momentum to be \begin{equation} p_t=\frac{\partial L}{\partial x'_t}(x_t,x'_t,t) \end{equation} From here, $x'_t$ can be determined as an implicit function $x'_t=\theta(x_t,p_t,t)$. The Hamiltonian is defined to be the Legendre transform \begin{equation} H(x_t,p_t,t)=\theta(x_t,p_t,t)\,p_t- L\left(x_t,\,\theta(x_t,p_t,t),\,t\right). \end{equation} Now notice that \begin{align} L(x_t,x'_t,t) &= \frac{d}{dt}J(x,t) = \frac{\partial J}{\partial t}(x,t) + x'_t\frac{\partial J}{\partial x}(x,t). \end{align}

For a stationary point $x$ of $J(-,t)$ it is satisfied that $\partial J/\partial x=\partial L/\partial x'_t=p_t$. Subtituting in the last equation and then in the definition of $H$ we get: \begin{equation} H(x_t,\frac{\partial J}{\partial x},t)+\frac{\partial J}{\partial t}(x,t) = 0, \end{equation} the Hamilton-Jacobi equation.

$\endgroup$

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.