Numerical Solutions of Singular Control Problems.
LOUISIANA STATE UNIV BATON ROUGE DEPT OF ELECTRICAL ENGINEERING
Pagination or Media Count:
In the thesis, three different methods are used in the optimization of control problems with the aid of digital computers. All three methods are extended to find singular solutions. The general technique used is to assume a nonoptimal control u sup m and compute a delta u which will drive the control closer to an optimal in order to minimize an integral cost function. The control must satisfy inequality constraints of the form ut or k, where k is a constant. The three methods used are optimum gradient, conjugate gradient, and parallel tangent. Two examples are presented with their respective analytical solutions and numerical solutions using each of the above methods. Comparisons are shown between the analytical and numerical solutions. Author
- Numerical Mathematics