Schr\"{o}dinger Equation Driven by the Square of a Gaussian Field: Instanton Analysis in the Large Amplification Limit

TL;DR

This paper analyzes the probability tail of the squared wave function amplitude driven by a Gaussian field, using instanton methods to reveal filamentary structures and divergence thresholds, supported by numerical simulations.

Contribution

It introduces the first instanton analysis of the Martin-Siggia-Rose action for this PDE, linking tail behavior to filamentary instantons and divergence thresholds.

Findings

Tail of $p(U)$ deduced from instanton statistics

Realizations of $S$ concentrate on filamentary instantons at large $ U$

Numerical simulations confirm bias towards instantons in large $ U$ samples

Abstract

We study the tail of $p(U)$, the probability distribution of $U=\vert\psi(0,L)\vert^2$, for $\ln U\gg 1$, $\psi(x,z)$ being the solution to $\partial_z\psi -\frac{i}{2m}\nabla_{\perp}^2 \psi =g\vert S\vert^2\, \psi$, where $S(x,z)$ is a complex Gaussian random field, $z$ and $x$ respectively are the axial and transverse coordinates, with $0\le z\le L$, and both $m\ne 0$ and $g>0$ are real parameters. We perform the first instanton analysis of the corresponding Martin-Siggia-Rose action, from which it is found that the realizations of $S$ concentrate onto long filamentary instantons, as $\ln U\to +\infty$. The tail of $p(U)$ is deduced from the statistics of the instantons. The value of $g$ above which $\langle U\rangle$ diverges coincides with the one obtained by the completely different approach developed in Mounaix et al. 2006 {\it Commun. Math. Phys.} {\bf 264}~741. Numerical simulations clearly show a statistical bias of $S$ towards the instanton for the largest sampled values of $\ln U$. The high maxima -- or `hot spots' -- of $\vert S(x,z)\vert^2$ for the biased realizations of $S$ tend to cluster in the instanton region.