30 juin 2020
30 juin 2020 de 10 h 00 à 11 h 00 (heure de Montréal/HNE) Réunion Zoom
In this talk we study the nonlinear Schrödinger equation posed on the 1-torus. Based on their numerics, Cho, Okamoto, & Shōji conjectured in their 2016 paper that: (C1) any singularity in the complex plane of time arising from inhomogeneous initial data is a branch singularity, and (C2) real initial data will exist globally in real time. If true, Conjecture 1 would suggest a strong incompatibility with the Painlevé property, a property closely associated with integrable systems. While Masuda proved (C1) in 1983 for close-to-constant initial data, a generalization to other initial data is not known. Using computer assisted proofs we establish a branch singularity in the complex plane of time for specific, large initial data which is not close-to-constant.
Concerning (C2), we demonstrate an open set of initial data which is homoclinic to the 0-homogeneous-equilibrium, proving (C2) for close-to-constant initial data. This proof is then extended to a broader class of nonlinear Schrödinger equation without gauge invariance, and then used to prove the non-existence of any real-analytic conserved quantities. Indeed, while numerical evidence suggests that most initial data is homoclinic to the 0-equilibrium, there is more than meets the eye. Using computer assisted proofs, we establish an infinite family of unstable nonhomogeneous equilibria, as well as heteroclinic orbits traveling between these nonhomogeneous equilibria and the 0-equilibrium.