Mathematical modelling of intellectual self-organizing automatic control system: action planning research

Mathematical modelling procedure is an integral part of complex system development process. Intellectual self-organizing automatic control systems are intended for functioning in conditions of changing the environment, controlled plant parameters, along with control purposes. Information incompleteness causes declarative statement of control task, i.e. without action sequence for its solution. As consequence, the major component of intellectual self-organizing automatic control systems is the action planning subsystem. Declarative tasks are solved by using artificial intelligence methods. However, existing methods of action planning represent the procedures demanding greater use of computing resources. Therefore efficiency of intellectual self-organizing automatic control systems in many respects is defined by productivity of action planning subsystem. Artificial neural planning networks are applied to increase efficiency as the mechanism of action planning in intellectual self-organizing automatic control systems. Mathematical modelling of intellectual self-organizing automatic control systems requires software realization of artificial neural planning networks. In this article we review the results of our study on the properties of artificial neural planning networks.