A study on the discovery of relevant fuzzy rules using pseudobacterial genetic algorithm

Abstract

This paper presents a new method for the discovery of relevant fuzzy rules using the Pseudo-Bacterial Genetic Algorithm (PBGA). The PBGA was proposed by the authors as a new approach combining a genetic algorithm with a local improvement mechanism inspired by a process in bacterial genetics, named bacterial operation. The presented system aims the improvement of the quality of the generated fuzzy rules, producing blocks of e ective rules and more compact rule bases. This is achieved by encoding the fuzzy rules in the chromosomes in a suitable form in order to make the bacterial operation more e ective and by using a crossover operation that adaptively decides the cutting points according to the distribution of degrees of truth values of the rules. In this paper, rst, results obtained when using the PBGA for a simple fuzzy modeling problem are presented and compared with other methods. Second, the PBGA is used in the design of a fuzzy logic controller for a semi-active suspension system. The results show the benets obtained with this approach in both of the studied cases.

Provision of quality of service for active services

“…Harvey [23] pointed out that gene interchange (as observed in bacteria [24,25]) could provide the rapid learning required. This was recently demonstrated by Furuhashi [26] for a bounded, globally optimised GA. In previous work [27] we have demonstrated that a novel unbounded, distributed GA with`b acterial learning'' is an eective adaptive control algorithm for the distribution of services in an active service provision system derived from the ALAN.…”

A game theoretic trust model for on-line distributed evolution of cooperation inMANETs

“…Different authors present them as microbial genetic algorithms (Harvey, 1996), bacterial algorithms (Cabrita et al, 2003), pseudobacterial genetic algorithms (Nawa et al, 1999), lateral gene transfer (Ochman et al, 2000) and plasmid migration (Marshall and Roadknight, 2000) among others. In general, these algorithms avoid the sexual reproduction, so they are not classified as parallel genetic algorithms, although they are extremely distributed.…”

“…The randomly chosen ith part of the mÀ 1 clones is mutated and the best fitted part is replicated in the m individuals. After this mutationevaluation-selection-replacement process is repeated for all the p parts, the fittest individual goes to the next population and the other mÀ1 individuals die (Nawa et al, 1999;Cabrita et al, 2003).…”

Elicitation and fine-tuning of fuzzy control rules using symbiotic evolution

“…Homaifar [10] used a GA to generate MFs and the rule-base simultaneously. Nawa and his co-workers [19] proposed a method for the discovery of relevant fuzzy rules using the pseudobacterial GA. However, in many of these methods the MFs partitions and the fuzzy labels are usually pre-deÿned and the rule base to be optimised becomes inevitably sizeable.…”