Evolving Rules Using Genetic Fuzzy Approach - an Educational Case Study


The work presents design and development of a system to automatically evolve rules through genetic-fuzzy  approach. The work highlights the advantages of genetic and fuzzy hybridization and proposes a framework to evolve rules automatically. The objective of the framework is to reduce the development effort and guide the development procedure in user  friendly fashion. The architecture presents a novel approach which integrates fuzzy logic with system design as well as back end of the framework is used to  evolve the fuzzy rules using genetic algorithm approach. To experiment the working of the proposed  framework, a system to measure multiple intelligence is selected. The paper describes the detail  methodology including encoding strategy of rules, suitable genetic operators and fitness function to evolve fuzzy rules for the selected system. The initial population of rules, evolved generations and output results are also described. The paper concludes with the scope and applications of the work to other domains.

By Rajendra Akerkar (co-authors: Kunjal Mankad, Priti S. Sajja)
In: International Journal on Soft Computing,  Vol.2,  No.1, pp 35-45, 2011,  AIRCC ISSN 2229 - 7103