Transfer of energy to high frequencies in the cubic defocusing nonlinear Schrödinger equation (Q977167): Difference between revisions

From MaRDI portal
RedirectionBot (talk | contribs)
Changed an Item
Import240304020342 (talk | contribs)
Set profile property.
Property / MaRDI profile type
 
Property / MaRDI profile type: MaRDI publication profile / rank
 
Normal rank

Revision as of 01:48, 5 March 2024

scientific article
Language Label Description Also known as
English
Transfer of energy to high frequencies in the cubic defocusing nonlinear Schrödinger equation
scientific article

    Statements

    Transfer of energy to high frequencies in the cubic defocusing nonlinear Schrödinger equation (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    21 June 2010
    0 references
    The authors treat the nonlinear Schrödinger equation: \(-i\partial_t u+\Delta u=|u|^2\), \(u(0,x)= u_0(x)\); \(x\in\mathbb{T}^2= \mathbb{R}^2/(2\pi\mathbb{Z})^2\), (torus). Two conservation laws for the solutions \(u\) of the Schrödinger equation hold: \[ E[u](t):=\int_{\mathbb{T}^2}(2^{-1}|\nabla u|^2+ 4^{-1}|u|^4)\,dx(t)= E[u](0) \] and \[ \int_{\mathbb{T}^2}|u|^2 \,dx(t)= \int_{\mathbb{T}^2}|u|^2\,dx(0), \] for all \(t> 0\). The Sobolev norms \[ \| u(t)\|_{H^s(\mathbb{T}^2)}:= \Biggl(\sum_{n\in \mathbb{Z}^2} \langle n\rangle^s|\widehat u(t, n)|\Biggr)^{1/2} \] are used, where \(\langle n\rangle:= (1+|n|^2)^{1/2}\), and \[ \widehat u(t,n) := \int_{\mathbb{T}^2}u(t,x)\exp(- in\cdot x)\,dx. \] Theorem. Let \(s> 1\), \(K\gg 1\), and \(0<\delta\ll 1\). There exists a global smooth solution \(u(t,x)\) of the Schrödinger equation, and a time \(T> 0\) with \(\| u_0\|_{H^s}\leq\delta\), and \(\| u(T)\|_{H^s}\geq K\). That is, \(H^s\) instability of the map \(u_0 \mapsto u\), \[ \inf_{\delta> 0} \Biggl(\limsup_{|t|\to\infty} \Biggl[\sup_{\| u_0\| H^s\leq\delta}\| u(t)\|_{H^s}\Biggr]\Biggr)= \infty, \] holds. The arguments used in the proof are: {\parindent=8mm \begin{itemize}\item[(I)] Let \(v(t,x):= \exp(iGt)u(t,x)= \sum_{n\in\mathbb{Z}^2} a_n(t)\exp\{i(n\cdot x+|n|^2 t)\}\). Then \[ -i\partial_t a_n= Ga_n+ \sum_{n_1- n_2- n} a_{n_1}\overline a_{n_3}\exp\{i\omega_4 t\}, \] where \(\omega_1= |n_1|^2- |n_2|^2+ |n_3|^2- |n|^2\). \item[(II)] By choosing \(G=-2\| u(t)\|_{L^2(\mathbb{T}^2)}\), \[ -i\partial_t a_n= -a_n|a_n|^2+ \sum_{n_1,n_2,n_3\in\Gamma(n)} a_{n_1}\overline a_{n_2} a_{n_3} \exp(i\omega_4 t), \tag{FNLS} \] is derived, \(\Gamma(n)= \{(n_1,n_2,n_3)\in (\mathbb{Z}^2)^3: n_1- n_2+ n_3= n\), \(n_1\neq n\), \(n_3\neq n\}\). \item[(III)] By using \(\Gamma_{\text{res}}(n)= \{(n_1, n_2, n_3)\in\Gamma(n): \omega_4=0\}\), the authors give resonant truncation of the FNLS, and use arguments related to the set of frequencies: \(\Lambda\subset\mathbb{Z}^2\). \end{itemize}}
    0 references

    Identifiers