Fuzzy Control of Trade-Off between Exploration and Exploitation Properties of Evolutionary Algorithms

In this paper novel selection method based on hybridization of fuzzy logic and previously elaborated mix selection is presented. Proposed method is based on fuzzy control of trade-off between exploration and exploitation properties of evolutionary algorithm. Two factors are considered in proposed selection method: the first is a number of generations of evolutionary algorithm, and second is a population diversity. Due to these two factors, we control the trade-off between global and local search of solution space, and also due to the fuzzy control the proposed method is more ”immune” in falling into the trap of local extremum. The results obtained using proposed method are compared with the results obtained using other selection methods like: roulette selection, elitist selection, fan selection, tournament selection, deterministic selection, truncation selection, and mix selection. The comparison is performed using test functions chosen from literature. The results obtained using proposed selection method are better in many cases than results obtained using other selection techniques.