Share Email Print
cover

Proceedings Paper

Iterative Solutions To Nonlinear Matrix Equations Using A Fixed Number Of Steps
Author(s): D. Casasent; A. Ghosh; C. P. Neuman
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

An iterative algorithm for the solution of a quadratic matrix equation (the algebraic Ricatti equation) is detailed. This algorithm is unique in that it allows the solution of a nonlinear matrix equation in a finite number of iterations to a desired accuracy. Theoretical rules for selection of the operation parameters and number of iterations required are advanced and simulation verification and quantitative performance on an error-free processor are provided. An error source model for an optical linear algebra processor is then advanced, analyzed and simulated to verify and quantify our performance guidelines. A comparison of iterative and direct solutions of linear algebraic equations is then provided. Experimental demonstrations on a laboratory optical linear algebra processor are included for final confirmation. Our theoretical results, error source treatment and guidelines are appropriate for digital systolic processor implementation and for digital-optical processor analysis.

Paper Details

Date Published: 28 November 1984
PDF: 7 pages
Proc. SPIE 0495, Real-Time Signal Processing VII, (28 November 1984); doi: 10.1117/12.944014
Show Author Affiliations
D. Casasent, Carnegie-Mellon University (United States)
A. Ghosh, AT&T Bell Laboratories (United States)
C. P. Neuman, Carnegie-Mellon University (United States)


Published in SPIE Proceedings Vol. 0495:
Real-Time Signal Processing VII
Keith Bromley, Editor(s)

© SPIE. Terms of Use
Back to Top