6
$\begingroup$

It is well-known that the following Hamiltonian system \begin{eqnarray} \left\{\begin{array}{rcl} \frac{dx}{dt}&=&y,\\ \frac{dy}{dt}&=&x(-1+x^2), \end{array}\right. \end{eqnarray} with $$ H(x,y)=\frac{x^2}2+\frac{y^2}{2}-\frac{x^4}{4}$$ has the solution $$ x=\tanh\left(\frac{t}{\sqrt2}\right),\qquad y=\frac1{\sqrt2}\text{sech}^2\left(\frac{t}{\sqrt2}\right) $$ such that $H(x,y)=\frac14$.

Now consider the following Hamiltonian system \begin{eqnarray} \left\{\begin{array}{rcl} \frac{dx}{dt}&=&y,\\ \frac{dy}{dt}&=&x(-1+x^4), \end{array}\right.\tag{1} \end{eqnarray} with $$ H(x,y)=\frac{x^2}2+\frac{y^2}{2}-\frac{x^6}{6}$$ and I want to find the explicit solution such that $H(x,y)=\frac{1}{3}$. I tried many ways to get the solution but failed. Any help is appreciated.

$\endgroup$
0

1 Answer 1

13
$\begingroup$

Edited.

Surprisingly, there is an elementary solution, if I made no mistake in the following computation.

Your equation is equivalent to $$\left(\frac{dx}{dt}\right)^2=\frac{1}{3}(x^6-3x^2+2),$$ (I just plugged $y=dx/dt$ to your Hamiltonian, and used its value $H=1/3$.) This equation is separable, $$\frac{t}{\sqrt{3}}=\int\frac{dx}{\sqrt{x^6-3x^2+2}}=:I$$ and requires inversion of the integral.

To reduce it to a standard integral, change $x^2=1/(u+1), \; dx=-(1/2)(u+1)^{-3/2}du,$ and the integral becomes $$I=-\frac{1}{2}\int\frac{du}{u\sqrt{2u+3}}.$$ The inverse function to this integral is elementary. Integrating and returning to the original variables, I obtained the general solution $$x(t)=\frac{2\sinh(t+c)}{\sqrt{6+4\sinh^2(t+c)}}.$$ Can you check this computation?

Of course, your value $1/3$ for the Hamiltonian is crucial here. With some other value you obtain a much more complicated integral which cannot be expressed in elementary functions. Where did this $1/3$ come from?

$\endgroup$
11
  • 2
    $\begingroup$ Thanks a lot. This is what I wanted. $\endgroup$ Commented Mar 28, 2019 at 21:29
  • $\begingroup$ Note it is much clearer than before. I really appreciate your help. $\endgroup$ Commented Mar 29, 2019 at 16:02
  • 1
    $\begingroup$ @xpaul: tell me where your $1/3$ comes from. $\endgroup$ Commented Mar 29, 2019 at 16:11
  • 1
    $\begingroup$ 1/3 is from fact that when $H(x,y)=\alpha$ passes through the separatrices $(\pm1,0)$, $\alpha=1/3$. $\endgroup$ Commented Mar 29, 2019 at 16:16
  • $\begingroup$ If I plug in your $x(t)$, I get $y=x'(t)=\frac{\sinh (c+2 t)}{\sqrt{\frac{\cosh (c+2 t)+1}{\cosh (c+2 t)+2}} (\cosh (c+2 t)+2)^2}$ but I find $\frac{x^2}2+\frac{y^2}{2}-\frac{x^6}6\neq\frac13$. $\endgroup$ Commented Mar 29, 2019 at 16:25

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.