An averaging theorem for nonlinear Schrödinger equations with small nonlinearities (Q476708)

In the paper, the averaging principle is established for the following nonlinear Schrödinger equation with a small perturbation \[ \frac{d}{dt}u + i(-\Delta u + V(x) u) = \epsilon \mathcal{P}(\Delta u, \nabla u, u, x), \quad x \in \mathbb{T}^d. \] Let \(\{ \xi_k \}_{k = 1}^{\infty}\) be the basis of eigenfunctions of the operator \(-\Delta u + V(x) u\), and let \(v_k\) be the Fourier coefficients of some function \(u\) with respect to this basis: \(u = \sum\limits_{k = 1}^{\infty} v_k \xi_k\). The map \(I(u) = \{ I_k(u)\}_{k = 1}^{\infty}\) is considered, where \(I_k(u) = \frac{1}{2} |v_k|^2\). The author proves that if the initial value problem for the NLS in the considered form has a unique solution \(u^{\epsilon}(t, x)\) for \(t \lesssim \epsilon^{-1}\) and some other conditions are fullfiled, then the sequence \(I(v^{\epsilon})\) converges as \(\epsilon \to 0\). Its limit \(I^0\) is a solution of the averaged equation, and it can be represented in the form \(I^0 = I(v)\), where \(v\) is the unique solution of the so-called effective equation.