Optimal Control of a Class of Reaction-Diffusion Systems with Pointwise State Constraints
In the authors' words
We study optimal control problems governed by a system of semilinear reaction-diffusion equations. The assumptions on the governing system are formulated so that our results apply to several models, including the Brusselator, Schnakenberg, and Gray--Scott systems. We prove the existence and uniqueness of a solution to the governing system without imposing any non-negativity assumption on the data. This allows us to study unconstrained, control-constrained, and state-constrained optimal control problems. For all these problems, we derive first-order necessary and second-order sufficient optimality conditions. The control acts on the activator equation through the boundary. Numerical examples demonstrate that both the inhibitor and the activator can be efficiently controlled.
Appeared: Wednesday, October 7. arXiv. Preprint, not yet peer-reviewed.
Authors' comment: 25 pages, 5 figures