OPTIMIZATION OF LARGE-SCALE DETERMINISTIC SYSTEMS USING DESCRIPTOR VARIABLE THEORY AND SPATIAL DYNAMIC PROGRAMMING Topical Report for March 1976-February 1978

摘要

Descriptor variable theory is a new approach for analyzing systems that preserves structure and allows considerable flexibility in modeling. This report summarizes the fundamental concepts used in descriptor analysis and some results concerning the structural properties of certain classes of descriptor systems. Optimization of a wide class of linear descriptor systems for quadratic performance criteria is discussed in detail. Spatial dynamic programming is a new approach to optimization that is particularly well-suited for large-scale systems by its exploitation of spatial structure. The procedure and its implications for large-scale systems design and con¬trol are described in the report.