Lanczos Algorithms for Large Symmetric Eigenvalue Computations Vol. I Theory (Progress in Scientific Computing)

CULLUM

ISBN 10: 146849192X / ISBN 13: 9781468491920

Published by Birkhaeuser, 2012

About this Item

reprint edition. 287 pages. 9.01x5.98x0.66 inches. In Stock.

Seller Inventory # x-146849192X

Bibliographic Details

• Title: Lanczos Algorithms for Large Symmetric Eigenvalue Computations Vol. I Theory (Progress in Scientific Computing)
• Publisher: Birkhaeuser
• Publication Date: 2012
• Language: English
• ISBN 10: 146849192X
• ISBN 13: 9781468491920
• Binding: Paperback
• Condition: Brand New
• Weight: 0.4 kilograms

About this title

Synopsis

Energy levels, resonanees, vibrations, feature extraetion, faetor analysis - the names vary from discipline to diseipline; however, all involve eigenvalue/eigenveetor eomputations. An engineer or physicist who is modeling a physieal proeess, strueture, or deviee is eonstrained to seleet a model for whieh the subsequently-required eomputations ean be performed. This eonstraint often leads to redueed order or redueed size models whieh may or may not preserve all of the important eharaeteristies of the system being modeled. Ideally, the modeler should not be foreed to make such apriori reduetions. It is our intention to provide here proeedures wh ich will allow the direct and suceessful solution of many large 'symmetrie' eigenvalue problems, so that at least in problems where the computations are of this type there will be no need for model reduetion. Matrix eigenelement eomputations can be c1assified as smalI, medium, or large seale, in terms of their relative degrees of difficulty as measured by the amount of computer storage and time required to eomplete the desired eomputations. A matrix eigenvalue problem is said to be sm all scale if the given matrix has order smaller than 100. Well-documented and reliable FORTRAN pro grams exist for small scale eigenelement computations, see in particular ElS- PACK [1976,1977]. Typically those programs explicitly trans form the given matrix into a simpler canonieal form. The eigenelement eomputations are then performed on the canonical form.