Controllability of quasi-linear Hamiltonian NLS equations (Q1678239)

In this paper the authors study the controllability in arbitrary time for a class of nonlinear NLS equations on the circle of the form \[ \partial_t u+i\partial_{xx}u+\mathcal N(x,u,\partial_x u,\partial_{xx}u)=0,\, x\in\mathbb R/2\pi\mathbb Z \tag{1} \] for the complex-valued unknown \(u=u(t,x)\), where \(\mathcal N\) is a Hamiltonian, quasi-linear nonlinearity. They prove an exact, internal controllability of the Cauchy problem \[ \begin{cases} \partial_t u+i\partial_{xx}u+\mathcal N(x,u,\partial_x u,\partial_{xx}u) =f\text{ for all }t\in [0,T]\text{ and }x\in\mathbb R/2\pi\mathbb Z,\\ u(0,x)=u_{\mathrm{in}}(x),\end{cases}\tag{2} \] with \[ u(T,x)=u_{\mathrm{end}}(x).\tag{3} \] Due to the presence of derivatives in the nonlinearity \(\mathcal N\), the controllability of the control problem (2)--(3) cannot be directly deduced by a perturbative argument. The authors prove the controllability of (2)--(3) by applying the Nash-Moser-Hörmander implicit function theorem of [the first and the second author, J. Funct. Anal. 273, No. 12, 3875--3900 (2017; Zbl 06792303)] as a black box. For this purpose, one needs to solve the associated linearized control problem, which is a real system with variable coeffcients. The authors solve this linearized problem by a reduction procedure to constant coefficients which is an adaptation of the procedure used in [\textit{R. Feola} and \textit{M. Procesi}, J. Differ. Equations 259, No. 7, 3389--3447 (2015; Zbl 1334.37087); ``KAM for quasi-linear forced Hamiltonian NLS'', Preprint, \url{arxiv:1602.01341}], and by applying the Hilbert uniqueness method.