OPTIMAL CONTROL AND CONVEX PROGRAMMING,
Abstract
A general type of discrete optimal control problem with both control and state constraints is considered. Necessary conditions for a relative minimum are given (assuming only differentiability) based on the KuhnTucker theory. For a convex function and linear system of differential equations it is shown that these conditions are also sufficient for a global minimum. A computational scheme is described for the state constrained problem where the conditions are sufficient. The scheme is based on a convex programming method and determines first if any admissible control exists, and if so, finds an optimal control. The solution of a four-dimensional system with state constraints is presented in order to illustrate this computational scheme. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1965
- Accession Number
- AD0614939
Entities
People
- J. B. Rosen
Organizations
- University of Wisconsin–Madison