Controllability of a $4\times4$ quadratic reaction-diffusion system

TL;DR

This paper proves local and global controllability of a 4x4 nonlinear reaction-diffusion system modeling chemical reactions, using linearization, Carleman estimates, and fixed-point arguments, applicable in any spatial dimension and time.

Contribution

It introduces a novel controllability analysis for a complex reaction-diffusion system with invariant quantities, extending control results to higher dimensions and large times.

Findings

Local exact controllability in any dimension and time

Global controllability in low dimensions for large times

Use of Carleman estimates and fixed-point methods

Abstract

We consider a $4\times4$ nonlinear reaction-diffusion system posed on a smooth domain $\Omega$ of $\mathbb{R}^N$ ($N \geq 1$) with controls localized in some arbitrary nonempty open subset $\omega$ of the domain $\Omega$. This system is a model for the evolution of concentrations in reversible chemical reactions. We prove the local exact controllability to stationary constant solutions of the underlying reaction-diffusion system for every $N \geq 1$ in any time $T >0$. A specificity of this control system is the existence of some invariant quantities in the nonlinear dynamics. The proof is based on a linearization which uses return method and an adequate change of variables that creates crossed diffusion which will be used as coupling terms of second order. The controllability properties of the linearized system are deduced from Carleman estimates. A Kakutani's fixed-point argument enables to go back to the nonlinear parabolic system. Then, we prove a global controllability result in large time for $1 \leq N \leq 2$ thanks to our local controllabillity result together with a known theorem on the asymptotics of the free nonlinear reaction diffusion system.