Artículo
The spatially variant fractional Laplacian
Fecha de publicación:
10/2023
Editorial:
Springer
Revista:
Fractional Calculus and Applied Analysis
ISSN:
1314-2224
Idioma:
Inglés
Tipo de recurso:
Artículo publicado
Clasificación temática:
Resumen
We introduce a definition of the fractional Laplacian (−Δ)^s(⋅) with spatially variable order s:Ω→[0,1] and study the solvability of the associated Poisson problem on a bounded domain Ω. The initial motivation arises from the extension results of Caffarelli and Silvestre, and Stinga and Torrea; however the analytical tools and approaches developed here are new. For instance, in some cases we allow the variable order s(⋅) to attain the values 0 and 1 leading to a framework on weighted Sobolev spaces with non-Muckenhoupt weights. Initially, and under minimal assumptions, the operator (−Δ)^s(⋅) is identified as the Lagrange multiplier corresponding to an optimization problem; and its domain is determined as a quotient space of weighted Sobolev spaces. The well-posedness of the associated Poisson problem is then obtained for data in the dual of this quotient space. Subsequently, two trace regularity results are established, allowing to partially characterize functions in the aforementioned quotient space whenever a Poincaré type inequality is available. Precise examples are provided where such inequality holds, and in this case the domain of the operator (−Δ)^s(⋅) is identified with a subset of a weighted Sobolev space with spatially variant smoothness s(⋅). The latter further allows to prove the well-posedness of the Poisson problem assuming functional regularity of the data.
Archivos asociados
Licencia
Identificadores
Colecciones
Articulos(IMAS)
Articulos de INSTITUTO DE INVESTIGACIONES MATEMATICAS "LUIS A. SANTALO"
Articulos de INSTITUTO DE INVESTIGACIONES MATEMATICAS "LUIS A. SANTALO"
Citación
Ceretani, Andrea Noemí; Rautenberg, Carlos N.; The spatially variant fractional Laplacian; Springer; Fractional Calculus and Applied Analysis; 26; 6; 10-2023; 2837-2873
Compartir
Altmétricas