Decomposition of Multivariable Systems for Distributed Fuzzy Control

Abstract

The control law (2.2) corresponds to a centralized consideration of the multivariable system (2.1) and it is evident that the whole number of fuzzy relations in this case will be n.m. However, this may be quite a great number and therefore time constraints in on-line control computations may be violated. On the other hand, an arbitrary decomposition of the system with the purpose of applying decentralized or hierarchical fuzzy control algorithms would not be always reliable. More specifically, a decentralized control structure may not guarantee good behavior of the system because of unaccounted interactions among subsystems and a hierarchical control structure may be very expensive because of the information transfer between the levels. Therefore, the system (2.1) must be decomposed in such a way that would guarantee successful application of distributed fuzzy control algorithms.