Knowledge-based function optimization using fuzzy culturalalgorithms with evolutionary programming

摘要

In this paper, the advantages of a fuzzy representation in problemnsolving and search is investigated using the framework of Culturalnalgorithms (CAs). Since all natural languages contain a fuzzy component,nthe natural question is “Does this fuzzy representation facilitatenthe problem-solving process, within these systems”. In order toninvestigate this question we use the CA framework of Reynolds (1996),nCAs are a computational model of cultural evolution derived from andnused to express basic anthropological models of culture and itsndevelopment. A mathematical model of a full fuzzy CA is developed there.nIn it, the problem solving knowledge is represented using a fuzzynframework. Several theoretical results concerning its properties arenpresented. The model is then applied to the solution of a set of 12ndifficult, benchmark problems in nonlinear real-valued functionnoptimization. The performance of the full fuzzy model is compared with 8nother fuzzy and crisp architectures. The results suggest that a fuzzynapproach can produce a statistically significant improvement in searchnefficiency over nonfuzzy versions for the entire set of functions, thenthen investigate the class of performance functions for which the fullnfuzzy system exhibits the greatest improvements over nonfuzzy systems.nIn general, these are functions which require some preliminaryninvestigation in order to embark on an effective search