Analisis Kinerja Min-Max Fitness Mutation pada Algoritma Genetika

Abstract

Implementation of method mutations in genetic algorithms that are not coordinated properly can make declining cause on an individual fitness value. To overcome these problems do settings for use mutation methods that are expected to make the process of finding a solution by genetic algorithms become more focused. In this study, the process of finding a solution using genetic algorithm is then performed on some of the individual with a minimum and maximum fitness value. Approach method used mutations are mutations min-max, where individuals with a minimum fitness value transferred in order to decrease and it is unlikely to survive, otherwise the individual with maximum fitness value is transferred in order to increase the survival chances so it can survive in the next process. In performance testing, a solution can be found quickly and the number of generations needed by process can to be reduce, but the restrictions on the number of generations is not always found a valid solution. Further testing with a small number of individuals show better performance than the larger number of individuals.