METHODS FOR OPTIMIZATION OF GENETIC ALGORITHMS CONSTRUCTING FINITE STATE MACHINES
Read the full article
Three ways is proposed to increase the rate of convergence of standard genetic algorithms for the class of problems in which the solution is a finite automaton. First of them allows to greatly reduce number of automaton states thanks to usage of flags. Method of recovering of links between states allows reducing number of unused states. Method of sorting states in usage order allows reducing number of states in automata.