Statistical theory of gradient systems of automatic optimization wish quadratic characteristic of object

The problem of analyzing gradient systems of automatic optimization with random factors is formulated. The methods of determining errors in the steady-state operation and the estimate of time of the transient process in the gradient system with many input variables are given. To illustrate the results the one-variable system and the two-variable system are considered.