A FIXED CONFIGURATION APPROACH TO STOCHASTIC CONTROL,

Abstract

A general two-point boundary-value problem (TPBVP) formulation is developed for a large class of stochastic optimization problems under a variety of control structure specifications. The application of the fixed-configuration technique to the stochastic linear regulator problem, and a class of stochastic nonlinear regulator problems is investigated. In the nonlinear case, analysis is not exact, and evaluation of the use of the method is with respect to the particular problem considered. The results obtained by investigating application of the fixed-configuration method to the stochastic linear regulator problem are of a more general nature. Some useful aspects of the approach are: (1) Alternate derivations of familiar optimal results are obtained. (2) It is shown that the optimal controller is not unique. (3) Simplified lower control structures are shown to be feasible. On the particular example investigated, near optimal results are obtained when a lower order control structure is used. (Author)

Document Details

Document Type
Technical Report
Publication Date
Jan 05, 1970
Accession Number
AD0705609

Entities

People

  • Craig S. Sims
  • James L. Melsa

Organizations

  • Southern Methodist University

Tags

DTIC Thesaurus Topics

  • Boundaries
  • Boundary Value Problems
  • Mathematics
  • Optimization
  • Regulators
  • Specifications
  • Stochastic Control
  • Test And Evaluation

Fields of Study

  • Engineering
  • Mathematics

Readers

  • Operations Research
  • Robotics and Automation.