Examining the Four Parameters of Genetic Algorithm in Order to Obtain the Best Solution for Transportation Network Design Problems

Usually, after carrying out network design studies, the question arises of how the output variables are affected by the input variables of a model. In other words, how can one use a method to change the inputs of a statistical model in an organized manner so that the effects of these changes can be predicted on the output of the model. This work is usually done by sensitivity analysis, so in this sensitivity analysis study, two parameters, the optimality degree of the objective function of the problem and the time to solve the problem, are examined in relation to the change in the four parameters of the genetic algorithm. In other words, the purpose of this study is to find the best combination of genetic algorithm parameters (survival probability, mutation rate, recombination probability, population) and also to check the changes of the mentioned criteria against the four parameters of the problem, so that the best solution is obtained. The results showed that increasing the initial population will improve the answer. This is because a more accurate search is performed with a larger number of solving factors in the feasible space. The solution time of the model also shows the same ۶۵% as the optimality of the search objective function following the best reproduction probability value. The higher the survival probability, the more chromosomes of the current generation will be transferred to the next generation without any operation, which will naturally reduce the solution time.

Keywords:

Genetic Algorithm , Transportation network , network design , optimization , Mutation