A parameter optimizer based on genetic algorithm for the simulation of carbonate facies

In this work we present a model for the dispersion of rocks in carbonate environments. The parameters used in this model are provided by the user who is simulating. However, defining values for such parameters is a time-consuming task and can generate inaccuracy in the simulation. To get around this, we propose a parameter optimizer based on genetic algorithm, in order to refine physical and categorical parameters used in the presented model. To condition the simulation we use an objective facies model that is considered as the optimal map, and we refine the initial parameters provided by the user through the fitness function that compares the simulated map and the objective. We present two experiments, considering the facies dispersion model applied in the Australian Great Barrier Reef. To analyze the convergence profiles of the genetic algorithm, we present a sensitivity analysis that helps in choosing the input values for the genetic operators. Finally, we briefly discuss the applicability of the optimizer in geological process simulation softwares.