3
$\begingroup$

Consider the periodic nonlinear Schrödinger equation $$-i \partial_t u + \Delta u = f(|u|)u, \qquad u=u(t,x) \in \mathbb{C}, \; t\in \mathbb{R}, \; x\in \mathbb{T}^n,$$ where $\mathbb{T}:= \mathbb{R}/\mathbb{Z}$, and $f$ is a smooth real function. This equation can be considered as the Hamiltonian equation which is induced by the Hamiltonian function \begin{equation} \label{ham} H(u) := \int_{\mathbb{T}^n} \left( \frac{1}{2} |\nabla u|^2 + F(|u|) \right)\, dx, \end{equation} where $F'(s)=sf(s)$, and by the symplectic form \begin{equation} \label{sympnls} \omega(u,v) := - \text{im} \int_{\mathbb{T}^n} u(x) \overline{v}(x)\, dx. \end{equation} This means that the Schrödinger equation can be written in the form $\partial_t u = X_H(u),$ where the "Hamiltonian vector field'' $X_H$ is formally defined by inserting $H$ and $\omega$ into the identity $\omega(X_H,\cdot) = - dH$. This is, kindly taken, from https://arxiv.org/pdf/1405.3200.pdf

So, my question is, I have read in some many places that the Schrödinger equation is integrable; does that means that we can find an (infinite) family of Poisson commuting integrals $F_1,F_2,\ldots$, also commuting with $H$? And if so, does somebody knows how to derive them or, at least, where can I find a place where these integrals of motion are derived?

$\endgroup$

1 Answer 1

5
$\begingroup$

I am a physicist, and I may be misinterpreting the question, but I suspect that you might be being led astray by nomenclature.

Generally, the Schrodinger equation does not have a complete set of simple integrals of motion - we cannot easily solve the Helium atom, for instance. Your equation is in some senses even more complicated than the Schrodinger equation, as you've introduced a nonlinearity through $f(|u|)$.

I suspect the confusion is from the following: the phrase "Nonlinear Schrodinger Equation" is used to describe a very specific nonlinear Schrodinger equation: $$-i \partial_t u + \Delta u = k |u|^2 u$$ with $n=1$ coordinates. This equation is indeed integrable. However, the general equation you wrote, $$-i \partial_t u + \Delta u = f(|u|)u, \qquad u=u(t,x) \in \mathbb{C}, \; t\in \mathbb{R}, \; x\in \mathbb{T}^n$$ is not generically integrable.

$\endgroup$
4
  • $\begingroup$ This is a very useful comment! In the integrable case you talked about, do you know how could I find the integrals of motion? I assume they may be obtained by some Lax pair construction, but how specifically? $\endgroup$
    – kvicente
    Commented Aug 11, 2023 at 1:11
  • 1
    $\begingroup$ @kvicente This book: iopscience.iop.org/book/mono/978-0-7503-2428-1 collects a lot of the relevant results. $\endgroup$
    – Buzz
    Commented Aug 11, 2023 at 1:31
  • 1
    $\begingroup$ @kvicente Buzz's book is a great resource that I didn't know existed - I hadn't realized there were nice solutions for larger $n$, though maybe that makes sense in hindsight. Maybe I skimmed the book a little fast, but I didn't see tables of integrals of motion. I know that the first five integrals of motion (and a recurrence relation to generate more) are listed in Zakharov and Shabat's paper at web.archive.org/web/20160325140414/http://jetp.ac.ru/cgi-bin/dn/… in section 8. However, I'm not sure whether that recurrence relation generates only a subset of IOM. $\endgroup$
    – user196574
    Commented Aug 11, 2023 at 7:01
  • $\begingroup$ Thanks very much! Both references look interesting. $\endgroup$
    – kvicente
    Commented Aug 11, 2023 at 17:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .