Peter Lax showed that the differential operators $$L=-6\partial_x^2-u,\quad B=-4\partial_x^3-u\partial_x-(1/2)u_x$$ fulfilling the Lax equation $$\dot{L}+[L,B]=0$$ is equivalent to the KdV equation $$u_t+uu_x+u_{xxx}=0.$$

For the harmonic oscillator one knows a Lax representation given by two two-by-two matrices.

But: Are there differential operators $L$ and $B$ which are equivalent to the harmonic oscillator equation $$\ddot{q}+\omega q=0$$ and fulfilling the Lax equation? If yes: How can one derive the form of these operators?

  • $\begingroup$ Do you mean that Lax proved $(\dot{L}+[L,B])u=0 \Leftrightarrow u_t+u u_x + u_{xxx}=0$? (in particular, the $x$ is taken to be independent from time? i.e. $\partial_t (\partial^2_x)=\partial^2_x\partial_t$ or not?) However this is a PDE while the harmonic oscillator equation is an ODE. $\endgroup$
    – yuggib
    May 7, 2015 at 7:46
  • $\begingroup$ Yes that is what I meant. Besides the fact that the harmonic oscillator is an ODE, is there no way to write down such matrices? Yes x is independent of time, so everything commutes. $\endgroup$
    – Richard
    May 7, 2015 at 7:52
  • 2
    $\begingroup$ Well, I am not sure if I understand completely, because $L=\partial_t$ and $B=\omega t$ (where $\omega\in\mathbb{R}$ does not depend on time) seem to do what you are asking...I will use the symbol $\partial$ here even if there is only $t$ as a variable, so it is not so meaningful (but shorter to write): $\partial_t L q=\partial_t^2 q$; $LBq=\omega q+\omega t\partial_t q$, $BLq=\omega t\partial_t q$, i.e. $[L,B]q=\omega q$. $\endgroup$
    – yuggib
    May 7, 2015 at 9:21

1 Answer 1


Well, consider a Lax pair for the Harmonic Oscillator; \begin{equation} L = \begin{pmatrix} p & \omega q \\ \omega q & -p \\ \end{pmatrix}, \quad M=\frac{\omega}{2} \begin{pmatrix} 0 & -1 \\ 1& 0 \\ \end{pmatrix} \end{equation} Since the Hamiltonian is $$H(q, p)=\frac{1}{2}(p^{2}+\omega^{2}q^{2})$$ It is eay to check that the Lax equation $\dot{L}=[M, L ]$ is equivalent to the Hamiltonian system \begin{equation} \dot{q}=\frac{\partial H}{\partial p}=p, \quad \dot{p}=-\frac{\partial H}{\partial q}=-\omega^{2}q \end{equation} It is a trivial exercise to show that \begin{equation} \ddot{q}+\omega^{2}q=0 \end{equation}


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.