A new scheme for an automatic generation of multi-variable fuzzy systems

We present a new novel method of automatically generating a multi-variable fuzzy inference system from the given sample sets. We first decompose the s ample set, say Lambda, into a cluster of sample sets associated with the gi ven input variables, then compute the associated fuzzy rules and membership functions for each variable, independent of the other variables, by solvin g a single input multiple output fUzzy system extracted from the set cluste r. The resulting decomposed fuzzy rules and membership functions for all th e variables are integrated back into the fuzzy system appropriate for the o riginal sample set ii. Taking advantage of the independence of the input va riables in computing the decomposed systems, we show that the computational complexity of the multi-variable system can in principle be reduced to tha t of a single variable if we can use a parallel processing multi-CPU system . We have verified our claim using an eight variable nonlinear function. (C ) 2001 Elsevier Science B.V. All rights reserved.