New Transformation Technique for Optimal Control Problems with Bounded State. Part I. Theory,

Abstract

The numerical solution of optimal control problems involving a state inequality constraint of the form L(x, theta) = or > 0 is considered. The approach employed is of the hybrid type, in an attempt to combine some of the best features of the indirect approach and the direct approach. While a predetermined number and sequence of subarcs are assumed (a feature of direct methods), enforcement of the state inequality constraint is obtained through a Valentine-type representation (a feature of indirect methods). A new transformation technique is developed by applying the Valentine-type representation to the kth derivative of the function L(x, theta), assumed to have a constant sign in each of the subarcs composing the extremal arc. The algorithm developed belongs to the class of sequential gradient-restoration algorithms. The principal property of the algorithm is that it produces a sequence of feasible suboptimal solutions: the functions x(t),u(t), pi obtained at the end of each cycle satisfy the constraints to a predetermined accuracy. Five numerical examples illustrating the theory are given in Part 2 (AD/A-004 334) and demonstrate the feasibility as well as the rapidity of convergence of the technique developed in this paper.

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1974
Accession Number
ADA004333

Entities

People

  • Angelo Miele
  • J. R. Cloutier

Organizations

  • Rice University

Tags

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Convergence
  • Inequalities
  • Sequences

Readers

  • Operations Research
  • Systems Analysis and Design