Synthesis of structure and algorithmic support of stochastic object management

Features, common for objects with the stochastic manner of operation, are described in the paper. The results are illustrated in order to give an example of objects of the considered class — a ball mill with dry cement grinding. The known approaches to the stochastic processes control are overviewed, and their advantages and deficiencies are mentioned. The modification of linear-quadratic-Gaussian (LQG) theory results is offered to be applied in the control loop. The modification is taking into account physical requirements to the calculation values of control actions by means of the application of limits to the range of their possible values. The Kalman filter modification is offered to be applied in the filter loop. It is based on the bias between the object output and the target values. The estimation technique of the proposed sub-optimal algorithms efficiency is developed. All the results, received during the research, are shown. The results prove the assumption about the efficiency of the offered variant of the algorithmic support.