Self-organizing fuzzy intelligent system

A self-organizing fuzzy system (SOFS) is presented. A plant model is not required for training, that is, the plant model is unknown to the SOFS. Using new data types, the vectors and matrices, a concise formulation is developed for the organization process and the inference activities of the SOFS. The fuzzy system can learn its rule-based structure and parameters from input/output training data. There is no fuzzy IF-THEN rule in the system initially. The fuzzy control policy is constructed automatically during the learning process when the system is simulated by input/output training data. With the well-known random optimization (RO) method, the generated fuzzy system can learn its parameters for specific applications. The proposed SOFS is applied to the temperature control problem.