Fractional derivatives have emerged as a powerful mathematical tool for modeling complex physical phenomena exhibiting memory effects and nonlocal interactions, which cannot be adequately captured by classical integer-order models. Fractional differential equations (FDEs) incorporating these derivatives provide a more accurate framework for describing anomalous diffusion, transport in heterogeneous media, and other processes with hereditary dynamics. However, the nonlocal nature that enhances their modeling capability also introduces substantial challenges, particularly in inverse problems where unknown sources, coefficients, or initial conditions must be recovered from indirect and often noisy measurements. Inverse problems for FDEs are typically ill-posed in the Hadamard sense, necessitating regularization techniques such as the quasi-boundary value method to restore stability and ensure meaningful reconstructions. The numerical solution of these problems relies on discretization techniques that produce large-scale linear systems whose coefficient matrices inherit the nonlocal structure of the fractional operators. Understanding the asymptotic spectral properties of these matrix sequences is essential for designing efficient solvers and preconditioning strategies. The theory of Generalized Locally Toeplitz (GLT) sequences provides a powerful and unified framework for this purpose, characterizing the spectral distribution of large structured matrices arising from discretizations and offering rigorous guidelines for preconditioner design. This thesis leverages GLT theory to analyze matrix sequences from inverse problems of FDEs, develop structure-preserving preconditioners, and investigate fundamental questions concerning nonlinear operations within the GLT algebra. In particular, it establishes maximal distribution results for geometric means of GLT sequences, proving that when the associated symbols commute, the geometric mean sequence remains GLT without requiring symbol invertibility, thereby resolving a long-standing conjecture while demonstrating that the result is maximal in the noncommuting degenerate case. Collectively, this thesis contributes to both the theory and numerical solution of inverse problems for FDEs, integrating regularization theory, spectral analysis, and preconditioning into a coherent computational framework. The theoretical advancements concerning GLT sequences and their geometric means further enrich the mathematical foundations of structured matrix analysis, providing practical strategies for solving inverse problems in complex physical systems governed by nonlocal dynamics
Fractional derivatives have emerged as a powerful mathematical tool for modeling complex physical phenomena exhibiting memory effects and nonlocal interactions, which cannot be adequately captured by classical integer-order models. Fractional differential equations (FDEs) incorporating these derivatives provide a more accurate framework for describing anomalous diffusion, transport in heterogeneous media, and other processes with hereditary dynamics. However, the nonlocal nature that enhances their modeling capability also introduces substantial challenges, particularly in inverse problems where unknown sources, coefficients, or initial conditions must be recovered from indirect and often noisy measurements. Inverse problems for FDEs are typically ill-posed in the Hadamard sense, necessitating regularization techniques such as the quasi-boundary value method to restore stability and ensure meaningful reconstructions. The numerical solution of these problems relies on discretization techniques that produce large-scale linear systems whose coefficient matrices inherit the nonlocal structure of the fractional operators. Understanding the asymptotic spectral properties of these matrix sequences is essential for designing efficient solvers and preconditioning strategies. The theory of Generalized Locally Toeplitz (GLT) sequences provides a powerful and unified framework for this purpose, characterizing the spectral distribution of large structured matrices arising from discretizations and offering rigorous guidelines for preconditioner design. This thesis leverages GLT theory to analyze matrix sequences from inverse problems of FDEs, develop structure-preserving preconditioners, and investigate fundamental questions concerning nonlinear operations within the GLT algebra. In particular, it establishes maximal distribution results for geometric means of GLT sequences, proving that when the associated symbols commute, the geometric mean sequence remains GLT without requiring symbol invertibility, thereby resolving a long-standing conjecture while demonstrating that the result is maximal in the noncommuting degenerate case. Collectively, this thesis contributes to both the theory and numerical solution of inverse problems for FDEs, integrating regularization theory, spectral analysis, and preconditioning into a coherent computational framework. The theoretical advancements concerning GLT sequences and their geometric means further enrich the mathematical foundations of structured matrix analysis, providing practical strategies for solving inverse problems in complex physical systems governed by nonlocal dynamics
Inverse problems with fractional operators: an analytical study and its numerical linear algebra treatment via the GLT theory / Asim Ilyas , 2026 Nov 11. 38. ciclo, Anno Accademico 2024/2025.
Inverse problems with fractional operators: an analytical study and its numerical linear algebra treatment via the GLT theory
ILYAS, ASIM
2026-11-11
Abstract
Fractional derivatives have emerged as a powerful mathematical tool for modeling complex physical phenomena exhibiting memory effects and nonlocal interactions, which cannot be adequately captured by classical integer-order models. Fractional differential equations (FDEs) incorporating these derivatives provide a more accurate framework for describing anomalous diffusion, transport in heterogeneous media, and other processes with hereditary dynamics. However, the nonlocal nature that enhances their modeling capability also introduces substantial challenges, particularly in inverse problems where unknown sources, coefficients, or initial conditions must be recovered from indirect and often noisy measurements. Inverse problems for FDEs are typically ill-posed in the Hadamard sense, necessitating regularization techniques such as the quasi-boundary value method to restore stability and ensure meaningful reconstructions. The numerical solution of these problems relies on discretization techniques that produce large-scale linear systems whose coefficient matrices inherit the nonlocal structure of the fractional operators. Understanding the asymptotic spectral properties of these matrix sequences is essential for designing efficient solvers and preconditioning strategies. The theory of Generalized Locally Toeplitz (GLT) sequences provides a powerful and unified framework for this purpose, characterizing the spectral distribution of large structured matrices arising from discretizations and offering rigorous guidelines for preconditioner design. This thesis leverages GLT theory to analyze matrix sequences from inverse problems of FDEs, develop structure-preserving preconditioners, and investigate fundamental questions concerning nonlinear operations within the GLT algebra. In particular, it establishes maximal distribution results for geometric means of GLT sequences, proving that when the associated symbols commute, the geometric mean sequence remains GLT without requiring symbol invertibility, thereby resolving a long-standing conjecture while demonstrating that the result is maximal in the noncommuting degenerate case. Collectively, this thesis contributes to both the theory and numerical solution of inverse problems for FDEs, integrating regularization theory, spectral analysis, and preconditioning into a coherent computational framework. The theoretical advancements concerning GLT sequences and their geometric means further enrich the mathematical foundations of structured matrix analysis, providing practical strategies for solving inverse problems in complex physical systems governed by nonlocal dynamics| File | Dimensione | Formato | |
|---|---|---|---|
|
master.pdf
accesso aperto
Descrizione: Inverse problems with fractional operators: an analytical study and its numerical linear algebra treatment via the GLT theory
Tipologia:
Tesi di dottorato
Dimensione
8.74 MB
Formato
Adobe PDF
|
8.74 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.



