On a long range segregation model

DOI10.4171/JEMS/747MaRDI QIDQ1687381zbMATH OpenOpenAlexFDO

Authors Luis Caffarelli, Stefania Patrizi, Veronica Quitalo

Publication date 29 December 2017

Published in Journal of the European Mathematical Society (JEMS) (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1505.05433

free boundary problems long-range interactions segregation of populations

Nonlinear elliptic equations (35J60) PDEs in connection with biology, chemistry and other natural sciences (35Q92) Free boundary problems for PDEs (35R35)

Abstract: In this work we study the properties of segregation processes modeled by a family of equations

L(u_i) (x) = u_i(x): F_i (u_1, ldots, u_K)(x)qquad i=1,ldots, K

where

F_{i} (u_{1}, l d o t s, u_{K}) (x)

is a non-local factor that takes into consideration the values of the functions

u_{j}

's in a full neighborhood of

x .

We consider as a model problem

Delta u_i^ep (x) = frac1{ep^2} u_i^ep (x)sum_{i eq j} H(u_j^ep)(x)

where

e p

is a small parameter and

H (u_{j}^{e} p) (x)

is for instance

H(u_j^ep)(x)= int_{mathcal{B}_1 (x)} u_j^ep (y), ext{d}y

or

H(u_j^ep)(x)= sup_{yin mathcal{B}_1(x)} u_j^ep (y).

Here the set

m a t h c a l B_{1} (x)

is the unit ball centered at

x

with respect to a smooth, uniformly convex norm

h o

of

e a l^{n}

. Heuristically, this will force the populations to stay at

h o

-distance 1, one from each other, as

e p o 0

.

Review text

The purpose of this work is a first attempt to study the properties of a segregation process. Basically, the authors consider a family of equations \[ L(u_i) (x) = u_i(x) F_i (u_1, \ldots, u_K)(x),\qquad i=1,\ldots, K, \] where \(F_i (u_1, \ldots, u_K)(x)\) is a non-local factor that takes into consideration the values of the functions \(u_i\) in a full neighborhood of \(x.\) They consider as a model problem the following system of equations \[ \begin{cases} \Delta u_i^\varepsilon (x) = \displaystyle\frac{1}{\varepsilon^2} u_i^\varepsilon (x) \sum\limits_{j\not= i} H(u_j^\varepsilon) (x) & \text{ in }\Omega, \\ u_i^\varepsilon=f_i & \text{ on }(\partial\Omega)_{\leq 1},\end{cases} \eqno{(2.4)} \] where \((\partial\Omega)_{\leq 1} := \left\{ x\in \Omega^c : d_\rho (x,\partial\Omega) \leq 1\right\}\), \(d_\rho (\cdot, K)=\inf_{y \in K} \rho(\cdot -y)\) is the distance function from \(K\) associated to a norm \(\rho\), and \(f_i\) are non-negative functions defined on \((\partial\Omega)_{\leq 1}\) with supports at \(\rho\)-distance \(\geq 1\) from each other: \[ d_\rho(\text{supp }\,f_i,\text{supp }\,f_j)\geq1\;,\quad \text{ for }\; i\neq j\,. \eqno{(2.3)} \] The authors denote \({\mathcal B}_r (x) := \left\{ y\in \mathbb{R}^n \mid \rho (x-y) < r\right\}\) and consider \[ H(w)(x) = \int_{{\mathcal B}_1(x)} w^p(y) \varphi \big(\rho (x-y)\big)\text{d}y,\qquad 1\leq p<\infty, \eqno{(2.5)} \] or \[ H(w) \big(x\big) = \sup_{{\mathcal B}_1(x)} w \eqno{(2.6)} \] with \(\varphi\) a strictly positive smooth function of \(\rho\), with at most polynomial decay at \(\partial{\mathcal B}_1\): \[ \varphi (\rho) \geq C(1-\rho)^q,\quad q\geq 0. \eqno{(2.7)} \] By a viscosity solution of the problem (2.4) the authors mean continuous functions \(u_1^\varepsilon,\ldots,u_K^\varepsilon\) that satisfy the system of equations (2.4) in the viscosity sense. Moreover, the authors assume: \[ \begin{cases} \varepsilon>0, \, \Omega \text{ is a bounded Lipschitz domain of } \mathbb{R}^n, \cr f_i: (\partial\Omega)_{\leq 1}\to\mathbb{R},\,f_i\geq 0,\,f_i\not\equiv0,\, f_i\text{ is Hölder continuous}, \cr \exists \,c>0 \,\forall x\in\partial\Omega\cap\text{supp}\,f_i,\, |{\mathcal B}_r(x)\cap \text{supp}\,f_i|\geq c|{\mathcal B}_r(x)|,\cr \text{(2.3) holds true},\cr H \text{ is either of the form (2.5) or (2.6) and (2.7) holds.}\end{cases} \eqno{(2.8)} \] The main result of this article is the following theorem. Theorem 4.1. Assume (2.8). Then there exist continuous positive functions \(u_1^\varepsilon,\ldots,u_K^\varepsilon\), depending on the parameter \(\varepsilon\), which form a viscosity solution of the problem (2.4). The authors also investigate the limit problem for \(\varepsilon\to 0\) and obtain several results about properties of the boundary sets \(\partial\{u_i>0\}\).

Recommendations

Cites work

Cited in

(14)

This page was built for publication: On a long range segregation model

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1687381)