Systematic design approach for multivariable fuzzy expert system
Bin-Da Liu, Chun-Yueh Huang · 1994
An algorithm to improve the Gupta's structure is proposed for the design of a multivariable fuzzy expert system. By this algorithm, the mapping problem is solved but the properties of systematic analysis and the regular design procedure are preserved. A new scheme which can save great amount of storage in implementing the relation matrices is presented. A famous example of inverted pendulum in fuzzy system is used to explain this algorithm. Taking the advantage of the composition rule of fuzzy inference, the inference speed of this structure is fast and it is suitable for real time processing.>