Analysis of Variance for A Influence of Genetic Probability on the Convergence times of Genetic Algorithm

摘要

The three-factor about replication, Exchange, mutation probabilities and their interaction is important on the convergence of the genetic algorithm. In order to analysis the effect. This study designed an three-factor cross-group experiment of three probability about reproduction, crossover and mutation, we use analysis of variance on empirical datum to obtain the effect of reproduction, crossover and mutation on convergence of Genetic algorithm. Conclusion of study show that, the influence of reproduction is the most significant, the influence of crossover is the second, the influence of mutation is not significant. But, as the main way of producing new individuality, Reproduction probability is important for getting overall situation optimal solution. In addition, the instrumental error is great which show that the calculate result of Genetic algorithm is randomness.