Automated and efficient order selection in Krylov-based model order reduction
by Mohammad Abid Bazaz; Mashuq-un-Nabi; S. Janardhanan
International Journal of Modelling, Identification and Control (IJMIC), Vol. 18, No. 4, 2013

Abstract: Krylov-subspace-based model order reduction is one of the most widely used techniques for the reduction of very large systems, having sizes ranging typically from thousands to several millions. By projecting the large system dynamics onto an appropriate Krylov subspace, some of the leading moments of the reduced system transfer function can be made to implicitly match the corresponding moments of the original system. The iterative algorithms used for generating the orthogonal bases for the Krylov subspace have reliable numerical implementations involving matrix-vector multiplications exclusively. However, the selection of a suitable order of the reduced system is an adhoc procedure in these algorithms. This paper presents an efficient stopping criterion based on a new index, known as coefficient of numerical rank improvement, to automate the process of order selection in Krylov-based reduction procedures. Numerical experiments conducted on some benchmark models show a significant reduction in the size of the reduced models as compared to those obtained through conventional procedures.

Online publication date: Sat, 16-Aug-2014

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Modelling, Identification and Control (IJMIC):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com