Fractional nonlinear degenerate diffusion equations on bounded domains part I. Existence, uniqueness and upper bounds. (January 2016)

• Record Type: Journal Article
• Authors: Bonforte, Matteo; Vázquez, Juan Luis
• Is Part Of: Nonlinear analysis. Volume 131(2016)
• Page Start: 363
• Page End: 398
• Publication Date: 2016-01
• Subjects: 35B45 -- 35B65 -- 35K55 -- 35K65; Nonlinear and nonlocal diffusion -- Fractional Laplacian on a bounded domain -- A priori estimates; Mathematical analysis -- Periodicals; Functional analysis -- Periodicals; Nonlinear theories -- Periodicals; Analyse mathématique -- Périodiques; Analyse fonctionnelle -- Périodiques; Théories non linéaires -- Périodiques; Functional analysis; Mathematical analysis; Nonlinear theories; Periodicals; Electronic journals; 515.7248
• DOI: 10.1016/j.na.2015.10.005
• Languages: English
• ISSNs: 0362-546X

Abstract

Abstract: We investigate quantitative properties of nonnegative solutions u ( t, x ) ≥ 0 to the nonlinear fractional diffusion equation, ∂ t u + L F ( u ) = 0 posed in a bounded domain, x ∈ Ω ⊂ R N, with appropriate homogeneous Dirichlet boundary conditions. As L we can use a quite general class of linear operators that includes the two most common versions of the fractional Laplacian ( − Δ ) s, 0 < s < 1, in a bounded domain with zero Dirichlet boundary conditions, but it also includes many other examples since our theory only needs some basic properties that are typical of "linear heat semigroups". The nonlinearity F is assumed to be increasing and is allowed to be degenerate, the prototype is the power case F ( u ) = | u | m − 1 u, with m > 1 . In this paper we propose a suitable class of solutions of the equation, and cover the basic theory: we prove existence, uniqueness of such solutions, and we establish upper bounds of two forms (absolute bounds and smoothing effects), as well as weighted- L 1 estimates. The class of solutions is very well suited for that work. The standard Laplacian case s = 1 is included and the linear case m = 1 can be recovered in the limit. In a companion paper (Bonforte and Vázquez, in preparation), we will complete the study with more advanced estimates, like the upper and lower boundary behaviour and Harnack inequalities, for which the results of this paper are needed.