| Summary: | 碩士 === 國立中興大學 === 電機工程學系所 === 95 === This paper proposes a self-generating fuzzy system with the learning ability coming from the proposal of an On-line Self-Aligning Clustering (OSAC) algorithm followed by Ant and Particle Swarm Cooperative Optimization (APSCO). The proposed OSAC algorithm not only helps to generate the rules from on-line training data but also helps to avoid the generation of highly overlapping fuzzy sets. Once a new rule is generated, the corresponding antecedent and consequent parameters are optimized by APSCO. In APSCO, ant colony and particle swarm coexist in a population, and they search an optimal parameter solution concurrently on each iteration. Ant paths not only help to determine the consequent parameters of generated rules but also help to generate auxiliary particles. Well-performed particles among the auxiliary particles and original particles are selected and the selected particles cooperate to find a better solution through particle swarm optimization. The proposed self-generating fuzzy system is applied to different fuzzy controller design problems. Comparisons with other evolutionary and swarm intelligence algorithms, and their hybrid are conducted to verify performance of the proposed method. The ACO used here is hardware implemented on Field Programmable Gate Array (FPGA) chip. The use of Programmable Logic Device (PLD) is more and more general in recent years, and the procedure of circuit deign is fast and elastic.
|
|---|
