Well-posedness and optimal control for a Cahn-Hilliard-Oono system with control in the mass term

TL;DR

This paper investigates the well-posedness and optimal control of a Cahn-Hilliard-Oono system with control in the mass term, addressing existence, regularity, and optimality conditions for different potentials in dimensions 1 to 3.

Contribution

It introduces a novel analysis of the control problem with general potentials, including singular ones, and establishes first-order optimality conditions under specific compatibility assumptions.

Findings

Proved well-posedness of the control system in dimensions 1 to 3.

Derived necessary optimality conditions for the control problem.

Addressed the strict separation property for logarithmic potentials in 2D.

Abstract

The paper treats the problem of optimal distributed control of a Cahn-Hilliard-Oono system in $\mathbb{R}^d$, $1\leq d\leq 3$, with the control located in the mass term and admitting general potentials that include both the case of a regular potential and the case of some singular potential. The first part of the paper is concerned with the dependence of the phase variable on the control variable. In the case of a logarithmic potential, we need to prove an ad hoc strict separation property, and for this reason we have to restrict ourselves to the case $d = 2$. In the rest of the work, we study the necessary first-order optimality conditions, which are proved under suitable compatibility conditions on the initial datum of the phase variable and the time derivative of the control, at least in case of potentials having unbounded domain. PLEASE NOTE: A revised version of this paper has been published in Discrete Contin. Dyn. Syst. Ser. S 15 (2022), 2135-2172: the authors point out that both the published version and the arXiv preprint contain a trivial mistake in formula (2.21), which leads to suitable changes in the assumption (2.10) and in some occurrences of it. Namely, (2.21) and (2.10) become $ \bar \varphi(t) = \bar \varphi_0 e^{-t} + \int_0^t e^{-(t-s)} \bar u(s) \, ds$ for every $t\in[0,T]$, and $-M - (\bar\varphi_0)^- , M + (\bar\varphi_0)^+$ belong to the interior of $D(\beta)$, respectively. The formula just after (2.21) becomes $-M - (\bar\varphi_0)^- \leq \bar \varphi(t) \leq M + (\bar\varphi_0)^+$ and a similar change must be done in (4.20). The first line of (2.24) is now implied by the new (2.21). Next, (2.39) becomes $-(M+ R) - (\bar\varphi_0)^- , (M+ R) + (\bar\varphi_0)^+$ belong to the interior of $D(\beta)$. Similar changes must be done in the interval written just two lines before (4.18) and in (4.18) itself.