Regularity Estimates for Fully Nonlinear Dead-Core Problems with a Hamiltonian Term

by, and, Ginaldo S. Sá · 2025

Read Regularity Estimates for Fully Nonlinear Dead-Core Problems with a Hamiltonian Term by by, and, Ginaldo S. Sá (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-2505-10778.txt.

Introduction

Rafael Costa and Ginaldo Sá study regularity for dead-core problems in this 2025 mathematics paper. The equations are fully nonlinear and elliptic, with a Hamiltonian term that can degenerate or blow up as the gradient nears zero. Non-negative solutions may vanish on whole regions, called plateau or dead-core zones. The free boundary, where the solution lifts off zero, is the focus. The authors prove that solutions grow away from that boundary in a regular way.

Under stated conditions on the exponents, the growth rate they find is optimal. A non-degeneracy result supports the sharpness claim. Two uses follow: solutions stay positive where expected, and the boundary of the coincidence set has controlled size in the Hausdorff sense. The work extends free boundary theory to operators with singular gradient terms.

Catalog record

CREATORby, and, Ginaldo S. Sá
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-10778

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Regularity Estimates for Fully Nonlinear Dead-Core Problems with a Hamiltonian Term Universidade Federal do Rio Grande do Sul, Porto Alegre-RS, Brazil E-mail address: [email protected] by, and, Ginaldo S. Sá Abstract In this paper, we present a problem involving fully nonlinear elliptic operators with Hamiltonian, which can present a singularity or degenerate as the gradient approaches the origin. The model studied here, allows the appearance of plateau zones, i.e. unknown regions of the domain in which the non-negative solutions vanishes. We show an improvement in regularity along the free boundary of the problem, and with some hypotheses on the exponents of the equation we proved the optimality of the growth rate with the help of the non-degeneracy also obtained here. In addition, some more applications of the growth results on the free boundary are obtained, such as: the positivity of solutions and also information on the Hausdorff measure of the boundary of the coincidence set. Keywords : Dead-core problem, improved regularity, fully nonlinear elliptic operator (degenerate/singular), Hamiltonian term. 2020 Mathematics Subject Classification: 35B65, 35J70, 35J75, 35R35 . 1 Introduction In this work, we investigate geometric properties of the viscosity solution to diffusion-reaction problems ruled by a uniformly elliptic operator with a Hamiltonian term, which may exhibit degenerate/singular behavior when the gradient vanishes, along with a strong absorption term: (1.1) |\nabla{u}|^{p}F(D^{2}u)+\mathfrak{a}(x)|\nabla u|^{q}=f(x,u)\lesssim\lambda_{0}(x)u(x)^{\mu}_{+}\quad\text{in}\quad\Omega, where \Omega\subset\mathbb{R}^{n} is a smooth, open and bounded domain, \mathfrak{a},\lambda_{0}\in\mathrm{C}^{0}(\overline{\Omega}) with \inf\lambda_{0}>0 , F:\text{Sym}(n)\rightarrow\mathbb{R} is a uniformly elliptic operator, where \text{Sym}(n) denote the space of symmetric matrices of size n\times n and • -1<p indicates the degree of singularity/degeneracy of the problem; • 0\leq q<p+1 sublinear exponent; • 0\leq\mu<p+1 represents the strong absorption term of the model. Additionally, we impose the boundary condition (1.2) u(x)=g(x)\quad\text{on }\partial\Omega, where the boundary datum is assumed to be continuous and non-negative, i.e., 0\leq g\in\mathrm{C}^{0}(\partial\Omega) . The equation approach ( 1.1 ) is classified as an equation with a strong absorption term, and the set \partial\{u>0\}\cap\Omega constitutes the free boundary of the problem. The study of such problems has an important role in the construction of abstract models that describe, for example, diffusion phenomena, where u represents the density of a chemical reagent or gas, while \lambda_{0} controls the ratio between the reaction rate and the diffusion rate. An illustrative example of an equation modeling the density of the reactant u in a stationary situation, see [ 2 , 5 , 6 ] for some motivations, is given by \left\{\begin{array}[]{rccccc}-\Delta u+\lambda_{0}f(u)\chi_{\{u>0\}}&=&0&\text{in}&\Omega,\\ u(x)&=&1&\text{on}&\partial\Omega,\end{array}\right. where \Omega\subset\mathbb{R}^{n} is a regular and bounded domain and f is a continuous and increasing reaction term, with f(0)=0 . In this context, f(u) represents the ratio of the reaction rate at concentration u to the reaction rate at concentration unity. From a mathematical perspective, these models are interesting due to the existence of a region where u\equiv 0 , called the dead-core . In the applied context, the dead-core corresponds to regions where the density of the substance (or gas) vanishes. Problems involving dead-core phenomena have been the subject of intense investigation in recent years. Results concerning the existence of solutions, estimates, and geometric properties, among other aspects, have been extensively developed (cf. [ 7 , 19 , 20 ] , just to cite a few). Another interesting aspect of the mathematical analysis of dead-core problems is its connection with various free boundary problems, see the seminal works [ 1 , 18 , 21 , 24 ] on problems with variational structure. See also [ 17 , 25 ] for the non-variational counterpart. The equation (1.1) has a singular/degenerate structure along the set \mathcal{C}(u)\coloneqq\{x\in\Omega:\,|\nabla u|=0\}. Such structure makes it difficult to obtain uniform regularity estimates, especially in neighborhoods of the set \mathcal{C}(u) , where the ellipticity of the operator deteriorates. In [ 22 ] , the authors proved interior Hölder regularity of the gradient for viscosity solutions in the following setting: (1.3) |\nabla u|^{\gamma}F(D^{2}u)=f(x)\in L^{\infty}(\Omega), where \gamma\geq 0 . In [ 3 ] , the authors obtained an optimal \mathrm{C}^{1,\alpha} -type estimate for the solutions of equation (1.1) along the set of critical points. In a sequence of celebrated works, Birindelli and Demengel addressed variants of singular/degenerate problems. In [ 9 ] , the authors considered singular/degenerate equations of the form (1.4) |\nabla u|^{\gamma}F(D^{2}u)=f(x,\nabla u), where F is fully nonlinear uniformly elliptic, \gamma>-1 , and f has, for the most part, growth of order 1+\gamma in the gradient term. In this work, they proved Lipschitz regularity results for viscosity solutions of (1.1). Furthermore, in [ 12 ] , the authors proved regularity \mathrm{C}^{1,\alpha} up to the boundary in the case \gamma\leq 0 , for the Dirichlet problem with homogeneous boundary conditions. In the context of fully nonlinear models of singular/degenerate type with Hamiltonian term, Birindelli and Demengel, in [ 13 ] , addressed the boundary regularity of viscosity solutions to the following problem (1.5) |\nabla u|^{p}F(D^{2}u)+|\nabla u|^{q}b(x)=f(x)\quad\text{in}\quad\Omega, where \Omega\subset\mathbb{R}^{n} is a bounded domain of class \mathrm{C}^{2} , p>-1 , q\in[0,p+1] , b,f\in\mathrm{C}(\overline{\Omega}) and F is a uniformly elliptic operator. In this context, the authors proved that given a boundary datum \varphi\in\mathrm{C}^{1,\gamma_{0}}(\partial\Omega) , there exist constants \gamma\in(0,\gamma_{0}) , with \gamma=\gamma(\lambda,\Lambda,p,q) and \mathrm{C}=\mathrm{C}(\gamma) such that any bounded viscosity solution u of \left\{\begin{array}[]{rrlcc}|\nabla u|^{p}F(D^{2}u)+|\nabla u|^{q}b(x)&=&f(x)&\text{in}&\Omega,\\ u&=&\varphi&\text{in}&\partial\Omega,\end{array}\right. is \mathrm{C}^{1,\gamma} and for p<q+1 \|u\|_{\mathrm{C}^{1,\gamma}(\overline{\Omega})}\leq\mathrm{C}\left(\|u\|_{\infty}+\|b\|_{\infty}^{\frac{1}{1+p-q}}+\|f\|_{\infty}^{\frac{1}{1+p}}+\|\varphi\|_{\mathrm{C}^{1,\gamma_{0}}(\partial\Omega)}\right). Moreover, for p=q+1 (linear case), the exponent \gamma also depends on \|b\|_{\infty} . We also emphasize the contributions of works [ 8 , 9 , 10 , 11 , 12 ] , which address fundamental results concerning the existence and uniqueness of solutions, maximum principles, and Harnack type inequalities. When the right-hand side of ( 1.1 ) is in L^{\infty} and F is a concave/convex operator, the best regularity one can expect is \mathrm{C}^{1,\frac{1}{1+p}}_{\operatorname{loc}} as p\geq 0 , see [ 3 , 22 ] . However, through an iterative geometric approach, our first result, see Theorem 1.1 , establishes a best pointwise regularity \mathrm{C}^{1,\beta} along the set of free boundary points \partial\{u>0\} , whenever 0<\mu<p+1 and 0<q<p+1 , since \beta=\min\left\{\frac{p+2-q}{p+1-q},\frac{p+2}{p+1-\mu}\right\}>1+\frac{1}{p+1}. In other words, even if we add a Hamiltonian term, we obtain a pointwise regularity at free boundary that is better than the local optimum estimate for degenerate uniformly elliptic equations with the convex operator, which is the content of Theorem 1.1 . It is important to note that in the paper [ 17 ] , the authors approach a similar analysis, the difference being that the drift term they deal with (\mathfrak{b}(x)\cdot\nabla u|\nabla u|^{p}) in an equation whose prototype is |\nabla u|^{p}F(D^{2}u,x)+\mathfrak{b}(x)\cdot\nabla u|\nabla u|^{p}=\mathfrak{\lambda}_{0}(x)u(x)^{\mu}_{+}, has an agreement with the rescaling of the entire equation. Our case can be seen as a continuation of this work, since when q=p+1 in ( 1.1 ), the Hamiltonian term has the same order as the drift term in [ 17 ] . In this sense, considering 0\leq q<p+1 brings significant difficulties, since the rescalings of the equation do not behave very well, which does not guarantee, for example, that the growth rate on the free boundary is optimal. However, the novelty of this work is to provide conditions to guarantee optimal growth on the free boundary through non-degeneracy, see Theorem 1.2 , which is not always valid depending on the exponents \mu and q . But, if \frac{p+2}{p+1-\mu}\leq\frac{p+2-q}{p+1-q}, the optimal growth rate along the boundary of

Read the complete text