DISCRETE TIME SYSTEM OPTIMIZATION ON ARBITRARY SETS AND FINITE DIMENSIONAL SPACES.
Abstract
As with continuous dynamical system optimization, the basic approaches to optimal control of discrete time systems have their origin in the variational theory of mechanics; Dynamical Programming stemming from the Hamilton-Jacobi Theory and the Discrete Minimum Principle based on the Hamilton Canonical Equations. In contrast to the usual heuristic and computational spirit associated with Dynamical Programming, we give precise necessary and sufficient conditions for optimality of discrete time systems over arbitrary sets, followed by some interesting observations concerning global successive approximations. This is followed by a review of the Discrete Minimum Principle as recently culminated by Halkin, Polak, and others; together with explicit connections between Bellman's equation and the discrete Canonical equations. A comprehensive development of the asymptotic properties of the discrete optimal linear regulator is then presented. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1969
- Accession Number
- AD0695074
Entities
People
- Mark H. Richardson
- R. Jeffrey Leake
Organizations
- University of Notre Dame