Comparative analysis of typical regulation algorithms and nonparametric dual control algorithm

The task of nonparametric dual control algorithm and standard regulation laws (P-, PI-, PID-regulators) at dy- namic objects with discrete-continuous nature of the technological process control is considered. It means that in itself, the dynamic process is continuous, however, the “input-output” variables of the process are controlled through discrete instants. In the case under study, the structure of the process model is partially parametrized. It means that equation order is determined from prior information, but at the same time functional dependency between “input-output” vari- ables of the process remains unknown. The methods of dynamic processes modeling and control based on nonparamet- ric algorithms are offered. The complexity of dynamic process modeling and control under condition of incomplete information is discussed. This level of prior information is characterised by the lack of model structure knowledge, but the information on object qualitative characteristics, for example, unambiguity, or ambiguity characteristics, linearity for dynamic processes or the nature of its nonlinearity is required. Methods of nonparametric statistics are applied to identification problem solving at this level of prior information. The problems of identification and control in the conditions of incomplete information are very relevant because many dynamic processes are not deeply studied and the presence of unknown distribution random noises causes more complexity in solving the identification and control tasks. The results of computing experiment which show the efficiency of nonparametric dual control algorithm in comparison with standard regulators are presented