Intelligent control: review.


The paper presents a review of the intellectual control systems: foundation, applications, program agents. The intellectual control is the most interesting application domain of theoretical cybernetics in the modern information systems. Here all developed receptions, methods, various parts of theory are going together: recognition of the image-states, concept formation, identification, optimization based on incomplete information, self-organizing automata theory. Practically all methods of operations research are joined here.
Area of practical application of methods of intellectual control is extraordinarily wide and includes: programs-agents for the difficult (especially functioning in the network environment) information systems; agents for the production workshop management; intellectual programs for decision-making in the regional systems; control programs in robotics; programs-researchers for automatic spaces vehicles and great deal other.
The dynamic intellectual systems, which able to rebuild itself as far as they are functioning in an environment being surrounding of controlled object, are most difficult. To realize such systems the design of mechanism of information "forgetting" is required for information which becomes in a time irrelevant.
Prime directions of development of intellectual control theory appear: development of specific models of memory for agents, algorithmic knowledge bases for agents, models of dynamic automatic state classification of the controlled objects and environment, automatic change of structures of mathematical models and even automatic changing of model in-use by an agent.