DYNAMIC AND HEURISTIC FUZZY CONNECTIVES-BASED CROSSOVER OPERATORS FORCONTROLLING THE DIVERSITY AND CONVERGENCE OF REAL-CODED GENETIC ALGORITHMS

Genetic algorithms are adaptive methods which may be used as approxima tion heuristic for search and optimization problems. Genetic algorithm s process a population of search space solutions with three operations : selection, crossover, and mutation. A great problem in the use of ge netic algorithms is the premature convergence, a premature stagnation of the search caused by the lack of diversity in the population and a disproportionate relationship between exploitation and exploration. Th e crossover operator is considered one of the most determinant element s for solving this problem. In this article we present two types of cr ossover operators based on fuzzy connectives for real-coded genetic al gorithms. The first type is designed to keep a suitable sequence betwe en the exploration and the exploitation along the genetic algorithm's run, the dynamic fuzzy connectives-based crossover operators, the seco nd, for generating offspring near to the best parents in order to offe r diversity or convergence in a profitable way, the heuristic fuzzy co nnectives-based crossover operators. We combine both crossover operato rs for designing dynamic heuristic fuzzy connectives-based crossover o perators that show a robust behavior. (C) 1996 John Wiley & Sons, Inc.