Some counterexamples concerning a model of eradication of invasive species

Abstract

In this note we study the regularity properties of the \emph{Effort Function $E(β)$}. This can be defined as the minimal $L^1$-norm of the control $\tilde α(x) \geq 0$ so that the PDE $$U_t = ΔU + f(U) - \tilde α(t,x) U, \quad U \in [0,1], %\ \tilde α\geq 0,$$ admits a traveling profile $U(x - βt)$ connecting $0$ to $1$. The above PDE models the evolution of an invasive species, where the control $\tilde α(t,x)$ describes the percentage amount of species removal. In the literature some conditions are assumed on the properties of $f(U)$ and (consequently) on the effort function $E(β)$, the main ones being that the traveling profile has a simple structure and that the effort function can be used as an anisotropic perimeter to generated a cost function over sets of finite perimeter. This note gives some examples showing that in the interesting cases, i.e. when the PDE has a non trivial dynamics, these assumptions are in general not satisfied. We also show that $E(β)$ is in general only $C^1$.