Towards hybrid soft computing approach to control of complex systems

Recently an approach to the control of complex systems employing a layered overall structure and different but compatible formalisms for subsystem representations on different levels, which is consistent with most of theoretical results in systems and control sciences, has been subject of extensive research. It provides a unified framework methodology for resolving system modeling identification and control design for complex multi-variable processes. One alternative of this approach is based on employing state space theory of composite similarity systems and the use of fuzzy systems, the other one makes use of neural networks instead, to deal with uncertainties and control adaptation. From the viewpoint of systems engineering, it may well be implemented within the standard computer process control technology.