A hybrid genetic algorithm with multi-parent crossover in fuzzy rule-based

Abstract

The fuzzy system has been widely used in several application fields and successfully performed by applying evolutionary. Genetic algorithm (GA) is one of the evolutionary methods for solving optimization problems. The success of GA depends on the design of its search operation which crossover and mutation are important operators to find a promising solution for difficult optimization problems. This article proposes a hybrid genetic algorithm with multi-parent crossover operators (HGA-MC) in fuzzy rule-based. An HGA-MC is used to optimize the fuzzy rule-based of linguistic values, which are associated with the global search. In experiments, the proposed algorithm and other existing algorithms were evaluated using optimization problems in UCI five datasets with different dimensionality. The experimental results showed that the proposed (fuzzy HGA-MC) achieved higher target precision than other existing methods by about 94.31%. Based on experimental results, HGA-MC could search for combinations of the crossover and mutation operators to discover accurate and concise optimization rules than other existing algorithms.