Optimal Linear Estimation and Suboptimal Numerical Solutions of Dynamical Systems Control Problems

摘要

A procedure for the generation of open loop suboptimal numerical solutions in the dynamical systems optimal control problem is proposed. The approximation of the control by a function dependent upon a finite number of parameters, and the use of a linear perturbation scheme associated with a direct search criterion reduce the problem to one of parameter optimization in each iteration. A stochastic approach to establish the search criterion and to treat the errors due to the first order approximation makes it possible to arrive at the search increment using optimal linear estimation.