Optimization of measurements in accelerated algorithms of extremal control (Q1820122)

The distribution of measurements of the output is considered for a plant with lag that is being optimized with constraints on the number of measurements and on the time interval between them. This can be done by minimizing the variance of the plant performance index with the use of a discrete extremal control system. Examples are presented of te distribution of measurements when a first-order plant is controlled by an algorithm that calculates the discontinuity of the first derivative and (respectively) by an algorithm that predicts the steady-state value. The results are compared in accordance with the probability of a false trial bias in the case of an optimal and of a uniform sampling respectively.