NEURAL  NETWORK   ALGORITHMS  BASED 
METHODS  FOR  ADAPTIVE  CONTROL  OF  COMPLEX  OBJECTS  BELONGING  TO  THE  CLASS
OF  NONLINEAR  DYNAMICAL  SYSTEMS

G. F. Malychina, A. V. Merkusheva

Saint-Petersburg

     Applied elements of the complex object control theory are considered when the object is characterized by many state parameters, nonlinear dynamics equations for parameters and the measurement system. The presen-tation, identification and control of such object (as a nonlinear multi-parametric dynamic system,  DS) based on the neural networks methodology are and described compared with those for linear DS. For the case of incom-plete information concerning the object model, the use of an adaptive control scheme is demonstrated.