The Rayleigh quotient iteration is a gem: starting with an approximate eigenvalue ( \mu ), solve ( (A-\mu I) y = x ), then update ( \mu ) to the Rayleigh quotient of ( y ). Parlett shows this converges cubically for symmetric matrices, but warns of pitfalls when near singular.
Any symmetric matrix can be diagonalized by an orthogonal matrix.
: Early chapters focus on methods where similarity transformations can be applied explicitly to the entire matrix.
One of the most powerful tools in the symmetric problem is the Rayleigh Quotient:
The book introduced many in the scientific community to the divide-and-conquer approach for finding eigenvalues of tridiagonal matrices, a technique that has become standard in modern high-performance computing libraries due to its parallel nature. The Impact on Modern Computing (LAPACK) parlett the symmetric eigenvalue problem pdf
: Though older, these methods are discussed for their reliability and potential for parallelization. Why This Work Matters
The book delves into advanced techniques like Cuppen’s divide-and-conquer method, which is highly efficient for large, parallelizable problems. E. Bisection and Inverse Iteration
Beresford Parlett’s "The Symmetric Eigenvalue Problem" is more than just a textbook; it is the definitive manual for anyone serious about computational mathematics. By balancing rigorous error analysis with practical algorithmic design, it remains as relevant today in the age of AI and big data as it was when first published in 1980.
One of the most important sections deals with how sensitive eigenvalues and eigenvectors are to changes in the matrix Abold cap A The Rayleigh quotient iteration is a gem: starting
The algorithms described by Parlett form the basis of LAPACK routines.
While the book is protected by copyright, it is widely recognized as a classic, and the "Classics in Applied Mathematics" edition is published by SIAM (Society for Industrial and Applied Mathematics).
Beresford N. Parlett’s The Symmetric Eigenvalue Problem is a seminal textbook in numerical analysis, not a single research paper. First published in 1980 by Prentice-Hall and later republished by the Society for Industrial and Applied Mathematics (SIAM) in their "Classics in Applied Mathematics" series, it serves as a comprehensive guide to the mathematics and algorithms behind computing eigenvalues and eigenvectors of real symmetric matrices. Google Books Summary of the Work
Parlett’s treatment of backward error and condition numbers for eigenvectors (via sin(Θ) theorems) is still sharper than most contemporary texts. : Early chapters focus on methods where similarity
Additionally, many university libraries provide access to the SIAM digital library. If you are a student or faculty member, check your institution's online catalog. Some library catalogs list access to an "electronic text (xxiv, 398 p.) : digital file" or a "Full Text (via SIAM)" option.
) of a real symmetric matrix are guaranteed to be real numbers.
) is critical for assessing how close an approximate eigenvalue is to the true spectrum of the matrix. Perturbation Theory and Rayleigh Quotients
: A deep dive into the workhorse of numerical linear algebra, showing how iterative matrix factorizations isolate eigenvalues.
8520 Dayton Pike - Soddy Daisy, TN 37379 - (423) 305-7494. Fax (423) 305-7496