Linear-quadratic optimization is one of the most exciting topics in the con trol engineering literature of recent decades. The interest in linear optimal dynamical systems, and especially in those with quadratic cost functional, can be explained by both the richness of properties these systems possess, and the pragmatic value and physical significance of the available results. Linear-quadratic optimization is not only a well-established theory, but also a powerful design tool, useful for time-invariant, as well as variable dynam ical systems.
Algorithms for Linear-Quadratic Optimization is intended to provide theoretical, algorithmic and computational support for solving the most frequently encountered linear-quadratic optimization problems. It is as self-contained as possible and includes both theoretical and practical topics. Theoretical treatment beneficially combines with an insight into algorith mic techniques, with relevant numerical issues, sophisticated software tools, implementation details that improve performance, practical recommenda- tions, and solution of specific problems. The algorithms are presented in a simple and concise informal language, a refined MATLAB notation, which facilitates computer implementation. Algorithmic templates are used rather than descriptions of source codes, to offer customization flexibility to both casual users and high-performance specialists, and to ensure the generality and independence of a specific computer language. Recent advances in nu merical linear algebra and associated mathematical software are taken into