|
|
When dealing with massive matrices (e.g., millions of rows and columns) arising from differential equations, dense methods fail due to memory limits. Parlett devotes significant coverage to the Lanczos algorithm. This method projects a large matrix onto a smaller Krylov subspace, extracting highly accurate approximations of the extreme eigenvalues (the largest and smallest) without altering the original sparse matrix structure. The Rayleigh Quotient and Error Bounds
Parlett, B. N. (1990). The symmetric eigenvalue problem. Prentice Hall.
The text is designed to provide the mathematical knowledge necessary for approximating eigenvalues and eigenvectors, particularly in the context of physical vibrations. It is structured into 15 chapters that progress from foundational theory to advanced computational techniques: Google Books Small to Medium Matrices (Chapters 1–9):
The book begins by establishing Basic Facts about Self-Adjoint Matrices , offering a quick but comprehensive tour of the linear algebra necessary for the discussions ahead. It immediately addresses the central question: What Is Small? What Is Large? in the context of computational cost and matrix density, setting the stage for understanding why certain algorithms are preferred.
Parlett demonstrates how the stationary points of this quotient correspond exactly to the eigenvectors of
While computing hardware has evolved from the mainframes of 1980 to modern distributed GPU clusters, the mathematical foundations detailed by Parlett have not changed. The algorithms described in his book form the backbone of modern numerical libraries like LAPACK, ARPACK, and MATLAB’s eig function.
This structure makes the book useful for a wide audience. Those new to the field can gain a solid foundation, while seasoned researchers will find deep insights into the most advanced computational strategies.
The Symmetric Eigenvalue Problem | SIAM Publications Library
See a comparing dense vs. tridiagonal solvers Share public link
(e.g., due to round-off errors). Parlett provides definitive bounds for perturbations, crucial for verifying the accuracy of numerical results. 3. Reduction to Tridiagonal Form
The book is a carefully structured journey, guiding the reader from fundamental concepts to sophisticated state-of-the-art methods. The chapter titles themselves convey the logical progression of the material, succinctly outlining the scope of the problem.
Because symmetric eigenvalue problems are inherently well-conditioned, Parlett shows that the computed eigenvalues are always close to the true eigenvalues of a slightly perturbed matrix. This mathematical guarantee underpins the reliability of modern numerical libraries like LAPACK and ARPACK. Finding Resources and Digital Copies
This article provides a comprehensive overview of the core concepts, algorithms, and mathematical insights found within Parlett’s classic text. 1. Why the Symmetric Eigenvalue Problem Matters In linear algebra, the eigenvalue problem seeks scalars and non-zero vectors such that: Ax=λxcap A x equals lambda x When the matrix is real and symmetric (
When dealing with massive matrices (e.g., millions of rows and columns) arising from differential equations, dense methods fail due to memory limits. Parlett devotes significant coverage to the Lanczos algorithm. This method projects a large matrix onto a smaller Krylov subspace, extracting highly accurate approximations of the extreme eigenvalues (the largest and smallest) without altering the original sparse matrix structure. The Rayleigh Quotient and Error Bounds
Parlett, B. N. (1990). The symmetric eigenvalue problem. Prentice Hall.
The text is designed to provide the mathematical knowledge necessary for approximating eigenvalues and eigenvectors, particularly in the context of physical vibrations. It is structured into 15 chapters that progress from foundational theory to advanced computational techniques: Google Books Small to Medium Matrices (Chapters 1–9):
The book begins by establishing Basic Facts about Self-Adjoint Matrices , offering a quick but comprehensive tour of the linear algebra necessary for the discussions ahead. It immediately addresses the central question: What Is Small? What Is Large? in the context of computational cost and matrix density, setting the stage for understanding why certain algorithms are preferred. parlett the symmetric eigenvalue problem pdf
Parlett demonstrates how the stationary points of this quotient correspond exactly to the eigenvectors of
While computing hardware has evolved from the mainframes of 1980 to modern distributed GPU clusters, the mathematical foundations detailed by Parlett have not changed. The algorithms described in his book form the backbone of modern numerical libraries like LAPACK, ARPACK, and MATLAB’s eig function.
This structure makes the book useful for a wide audience. Those new to the field can gain a solid foundation, while seasoned researchers will find deep insights into the most advanced computational strategies. When dealing with massive matrices (e
The Symmetric Eigenvalue Problem | SIAM Publications Library
See a comparing dense vs. tridiagonal solvers Share public link
(e.g., due to round-off errors). Parlett provides definitive bounds for perturbations, crucial for verifying the accuracy of numerical results. 3. Reduction to Tridiagonal Form The Rayleigh Quotient and Error Bounds Parlett, B
The book is a carefully structured journey, guiding the reader from fundamental concepts to sophisticated state-of-the-art methods. The chapter titles themselves convey the logical progression of the material, succinctly outlining the scope of the problem.
Because symmetric eigenvalue problems are inherently well-conditioned, Parlett shows that the computed eigenvalues are always close to the true eigenvalues of a slightly perturbed matrix. This mathematical guarantee underpins the reliability of modern numerical libraries like LAPACK and ARPACK. Finding Resources and Digital Copies
This article provides a comprehensive overview of the core concepts, algorithms, and mathematical insights found within Parlett’s classic text. 1. Why the Symmetric Eigenvalue Problem Matters In linear algebra, the eigenvalue problem seeks scalars and non-zero vectors such that: Ax=λxcap A x equals lambda x When the matrix is real and symmetric (
