Class of suboptimal feedback policies for linear and nonlinear systems

A class of suboptimal feedback-control policies which are easily on-line implemented and which give a value for performance index close to the truly optimal is suggested. Basic necessary conditions are derived for a general nonlinear case and specifically applied to linear quadratic problems. Numerical computations are made for linear and nonlinear examples and the results are tabulated and discussed. It is seen that, particularly in the case of nonlinear systems, an integral-state feedback introduced in the paper not only synthesises the control in a completely feedback configuration, but also improves system optimisation to a point where it is remarkably close to the truly optimal open loop.