Parameter Optimization of Genetic Algorithms for the Allocation of Highway Maintenance Resources

To reveal the influence of genetic algorithm parameters on the calculation efficiency and accuracy,parameters optimization based on orthogonal experiments is proposed for optimal allocation of highway maintenance resources.Three parameters,including the ratio of population size to iteration value,crossover form and mutation probability have been selected for three levels of orthogonal experiment.Statistics and critical values are calculated based on the hypothesis test models of the 3 parameters' main effects and interactive effects.Experimental results showed that the ratio of population size to iteration value and crossover form had the more significant influence on calculation efficiency and accuracy compared with the mutation probability,and that the optimal results of the 3 parameters being 1/4,two-point crossover,and 0.05.The method proposed reduces time in experiment for parameters estimation using genetic algorithms.