On a fractional semilinear Neumann problem arising in Chemotaxis
Eleonora Cinti, Matteo Talluri · 2025
Read On a fractional semilinear Neumann problem arising in Chemotaxis by Eleonora Cinti, Matteo Talluri (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-12181.txt.
Introduction
Chemotaxis, the drift of cells toward a chemical signal, is often modeled with equations that allow a spread-out, nonlocal kind of diffusion. Eleonora Cinti and Matteo Talluri study the steady patterns of such a model in a 2025 paper. Their problem is the fractional version of a classic one first studied in the 1980s, with no flow of the chemical across the boundary.
They extend earlier results, proved for one special fractional power, to every power between zero and one. When the diffusion strength is small, they show that non-constant steady patterns exist and describe some of their features. When diffusion is strong, every steady state must be constant. The results sharpen what mathematics can say about pattern formation in this model.
Catalog record
| CREATOR | Eleonora Cinti, Matteo Talluri |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 64 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-12181
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On a fractional semilinear Neumann problem arising in Chemotaxis Eleonora Cinti, Matteo Talluri Abstract . We study a semilinear and nonlocal Neumann problem, which is the fractional analogue of the problem considered by Lin–Ni–Takagi in the ’80s. The model under consideration arises in the description of stationary configurations of the Keller–Segel model for chemotaxis, when a nonlocal diffusion for the concentration of the chemical is considered. In particular, we extend to any fractional power s ∈ ( 0 , 1 ) s\in(0,1) of the Laplacian (with homogeneous Neumann boundary conditions) the results obtained in [ 20 ] for s = 1 / 2 s=1/2 . We prove existence and some qualitative properties of non–constant solutions when the diffusion parameter ε \varepsilon is small enough, and on the other hand, we show that for ε \varepsilon large enough any solution must be necessarily constant. 2010 Mathematics Subject Classification 35R11, 35B09, 35B45, 35A15, 60G22. 1. Introduction In this paper we consider the following fractional semilinear Neumann problem on a smooth and bounded domain \Omega\subset\mathbb{R}^{n} , (1.1) \begin{cases}(-\varepsilon\Delta_{N})^{s}u+u=u^{p}\quad&\text{on $\Omega,$}\\ \partial_{\nu}u=0\quad&\text{in $\partial\Omega,$}\\ u>0&\text{in $\Omega$},\end{cases} where \varepsilon>0, s\in(0,1) , 1<p<\frac{n+2s}{n-2s} , \partial_{\nu} denote the outward normal derivative to \partial\Omega and (-\Delta_{N})^{s} denotes the Spectral fractional Laplacian. This problem can be seen as a possible nonlocal analogue of the one studied by Lin, Ni, and Takagi in a series of seminal papers in the eighties ( [ 13 , 16 ] ). The case s=1/2 was considered by Stinga and Volzone in [ 20 ] , here we extend their results to any fractional power s\in(0,1) of the Neumann Laplacian. There are different possible generalization of the classical local problem to the non local setting, depending on which notion of fractional Laplacian and of Neumann boundary condition one considers. As already mention, here we focuse on the spectral Neumann Laplacian , as done in [ 20 ] , since this is interesting in applications. Thanks to the results in [ 18 , 19 ] , in order to study problem ( 1.1 ), we will study the associated local extended problem: (1.2) \begin{cases}\varepsilon\Delta_{x}U+\frac{1-2s}{y}U_{y}+U_{yy}=0&\text{ in }\mathcal{C},\\ \partial_{\nu}U=0&\text{ on }\partial_{L}\mathcal{C},\\ U(x,0)>0&\text{ in }\Omega\\ -\lim_{y\to 0}y^{1-2s}U_{y}(x,y)+U(x,0)=U(x,0)^{p}&\text{ in }\Omega,\end{cases} where \mathcal{C}:=\Omega\times(0,+\infty) and \partial_{L}\mathcal{C}:=\partial\Omega\times(0,+\infty) . Notice that here, just for convenience, we have preferred to put the parameter \varepsilon as a coefficient in the equation, rather than on the weighted normal derivative on \{y=0\} , see Remark 2.3 . Let a:=1-2s. Problem ( 1.1 ) has a variational structure and the associated energy functional is given by \begin{split}J_{\varepsilon,a}(U):=\frac{1}{2}\iint_{\mathcal{C}}\left(\varepsilon|\nabla_{x}U|^{2}+|U_{y}|^{2}\right)y^{a}dxdy+\int_{\Omega}|U(x,0)|^{2}dx-\frac{1}{p+1}\int_{\Omega}|U(x,0)|^{p+1}dx\end{split}. Our main results concern exitence and non–existence of non-constant solutions for problem ( 1.2 ), depending on the parameter \varepsilon . More precisely, we have the following: Theorem 1.1 . There exists \overline{\varepsilon}>0 such that if \varepsilon<\overline{\varepsilon} , then problem ( 1.1 ) admits at least one non–constant positive weak solution u^{\varepsilon,a} . Moreover, if we denote by U^{\varepsilon,a} its extension satisfying ( 1.2 ), then we have the following energy estimate: (1.3) J_{\varepsilon,a}(U^{\varepsilon,a})\leq C(n,a,p,\Omega)\varepsilon^{\frac{n}{2}}. For the notion of weak solution, see Section 4. In Section 6, we also prove some properties of such solutions, in particular we show that they tend to zero in measure and concentrates around some points as \varepsilon\rightarrow 0 . On the other hand, we show non–existence of non–constant solutions for sufficiently large \varepsilon , according to the following result Theorem 1.2 . There exists \varepsilon^{*}>0 such that if \varepsilon>\varepsilon^{*} then the unique weak solution of problem ( 1.1 ) is u\equiv 1. To prove existence of a weak solution to ( 1.2 ) (and thus of ( 1.1 )), we make a standard use of the Mountain Pass Theorem, then for showing that such solution cannot be constant if \varepsilon is small enough, the energy estimate ( 1.3 ) (proven by means of a suitable competitor) will be crucial. On the other hand, the proof of Theorem 1.2 crucially relies on some a priori (uniform in \varepsilon ) L^{\infty} -estimates for weak solutions, which will be obtained by a blow-up technique à la Gidas-Spruck (see Section 5). More precisely, we have the following Theorem 1.3 . Let u be a weak solution of ( 1.1 ). Then, there exists a positive constant C>0 , that does not depend on \varepsilon , such that \sup_{x\in\Omega}u(x)\leq C. We conclude this Introduction, describing the Keller-Segel model and how problem ( 1.1 ) arises in the study of Chemotaxis. Chemotaxis is the oriented movement of cells towards higher concentrations of some chemical substance in their environment. Keller and Segel proposed the following model to describe the chemotactic aggregation. Let w(x,t)>0 be the density of a bacteria population and v(x,t)>0 the density of a certain chemoattractant, then we have \begin{cases}\partial_{t}w=D_{1}\Delta w-\chi\text{div}(w\nabla(\log v))&\mbox{in }\Omega\times(0,+\infty)\\ \partial_{t}v=D_{2}\Delta v-aw+bv&\mbox{in }\Omega\times(0,+\infty)\\ \partial_{\nu}w=\partial_{\nu}v=0&\mbox{on }\partial\Omega.\par\end{cases} In [ 13 , 16 ] , Lin, Ni, and Takagi studied the steady states of the above system, relying on the observation that it could be reduced to the study of a single equation. Indeed, we can rewrite the stationary equation D_{1}\Delta w-\chi\text{div}(w\nabla(\log v))=0 as \text{div}(D_{1}w\nabla(\log w-\chi D_{1}^{-1}\log v))=0 and use the 0 –Neumann condition, to see that w=\lambda v^{{\chi/D_{1}}} for some positive \lambda . Hence the stationary version of the Keller–Segel system reduces to a single equation for v , which can be written in the form \begin{cases}-\varepsilon\Delta u+u=u^{p}&\text{in }\Omega\\ \partial_{\nu}u=0&\text{on }\partial\Omega,\par\end{cases} where p=\chi/D_{1} and \varepsilon=D_{2}/a . Possibly different nonlocal approaches to the study of Chemotaxis have been proposed recently in several papers. In [ 7 ] , Escudero proposed a model in which the chemical satisfies a classical local diffusion equation while the bacteria satisfy a nonlocal diffusion. In contrast, Stinga and Volzone [ 20 ] , considered a model in which the diffusion is assumed to be local for the bacteria and nonlocal for the chemical. As already mentioned, in this work we use the same approach, generalizing their results to any fractional power s\in(0,1) . Another nonlocal model for chemotaxis was proposed by Huaroto and Neves in [ 11 ] . They considered a fractional variant of the Keller-Segel model, in which the velocity (that in the classical model is proportional to the gradient of the concentration of the chemoattractant \nabla v ), is instead defined thorugh a linear fractional operator. This leads again to an equation for the chemical, in which the diffusion is governed by a fractional Neumann Laplacian. The paper is organized as follows: • In Section 2, we introduce the functional setting and the extension problem for the spectral Neumann fractional Laplacian; • Section 3 deals with regularity results for weak solutions; • In Section 4, we prove the existence result Theorem 1.1 ; • In Section 5, we prove the uniform L^{\infty} estimate of Theorem 1.3 and, as a consequence, the non–existence result Theorem 1.2 ; • Finally, Section 6 contains some properties of the non–constant solutions given by Theorem 1.1 . In particular, we show that they concentrate at some points (see Theorem 6.6 ). A crucial ingredient will be an Harnack inequality contained in [ 3 ] . Notations: Let \mathbb{R}^{n+1}=\{(x,y)\in\mathbb{R}^{n}\times\mathbb{R}\} and \mathbb{R}_{+}^{n+1}=\{(x,y)\in\mathbb{R}^{n+1}\,:\,y>0\} . We denote by B(x_{0},r) the ball of radius r centered at x_{0} in \mathbb{R}^{n} , and we set B_{r}:=B(0,r) . Moreover, \widetilde{B}((x_{0},y_{0}),r) denotes the ball of radius r