ALGORITHM OF 2-D SYMMETRIC FACTORIZATION

ZEZULA, P.; JEŽEK, J.

Abstract

The symmetric factorization is a mathematical and computional procedure constituting a key part of the linear-quadratic optimization respecting causality. This method of optimization dates back to Wiener, for filtering radar signals. Later it was broadly used in data processing and control. For 1-D systems, several algorithms for symmetric factorization are known, the numerical experiences are good. This paper is a 2-D modification of the algorithm. The generalization from 1-D to 2-D is by no means straightforward, new features are emerging. The 2-D LQ optimization is described here for a typical control problem. The problem is intentionally selected as simple as possible, to be able to concentrate to the main things.

Coresponding author e-mail: xzezulap[at]lab[dot]felk[dot]cvut[dot]cz

Session: Algorithms and Computing for Control