A Biologicaly Inspired Decision Model for Multivariable Genetic- Fuzzy-AHP System

This paper describes a hybridized intelligent algorithm as a tuning mechanism for one type of Genetic Fuzzy system termed the Genetic Fuzzimetric Technique (GFT). The proposed technique is based on the genetically inspired operations of crossover and mutation to achieve an optimized solution used to tune the fuzzy set shape (variables) within the rule-set. The GFT deals with knowledge representation in a modular form where each module -- termed a chromosome, in this article -- represents the defuzzified value of a rule-set inferring a specific output from a fuzzy input. A multivariable system, in this case, is the combination of all these chromosomes via a weighting factor termed the “Input Importance Factor”. This paper also explains the Analytic Hierarchy Process (AHP) technique which is proposed as a pairwise comparison methodology for the creation, selection and adaptation of the Input Importance Factor. The proposed GFT mechanism can be applied to any decision making problem within an uncertain environment. One example would be the determination of CRM performance measurement given a variety of inputs related to marketing, data mining tools, ordered materials and communication.