Hybrid and adaptive genetic fuzzy clustering algorithms

Abstract

This thesis proposes several effective clustering algorithms mainly based on genetic algorithms (GAs). A genetically guided clustering approach using an adaptive GA is proposed. The dynamic population size and varying crossover and mutation probabilities during the evolutionary process improve the convergence speed and convergence performance. To overcome the drawbacks of slow convergence speed in conventional GA, a micro-GA is applied instead of GA in the proposed algorithms. The performance of micro-GA is further improved by integrating with GA and simulated annealing (SA) in the two proposed hybrid genetic algorithms MGA and GAS. The use of GA or SA not only introduces new members into the population of micro-GA, but also a??leadsa?? micro-GA to evolve to good development by systematic simulated annealing process. The effectiveness of the proposed algorithms in clustering optimization is illustrated by simulation examples.