An Iterative Procedure for Optimal Control of a System by Sampled Input

Abstract
Optimal control of a given system by an amplitude constrained sampled input is achieved by using an iterative procedure. The optimality is based on the minimization of a function ⊘ of the state x(tf), tf being the end of the time over which control is applied. The iterative procedure is based on taking the F-derivative )Todd 1962) of ⊘ when a small change in state and input control is assumed. A steepest descent procedure based on p-norm in control space is used to produce a small δ⊘ for a small change in control. This change δ for minimization of ⊘ is made negative to decrease ⊘ and the procedure is now iteratively continued till no significant improvement in ⊘ is possible without violating a constraint. This then specifies the input control over given (to tf). If tf is varied the minimum time problem can also be solved. Computational technique as well as examples are given.

This publication has 4 references indexed in Scilit: