Multi-criteria decision aid is a research field aimed at providing decision-makers with tools to assist in decision-making based on multivariate data. Among these tools, the Ranking with Multiple Profiles (RMP) model allows for ordering a set of alternatives by pairwise comparison to reference profiles. The optimal configuration of such a model can be complex due to multiple parameters involved. Research has shown the usefulness of genetic algorithms in determining the parameters of this model. These algorithms aim to evolve a population of solutions through a series of crossover and mutation operators between parent solutions to produce child solutions. However, genetic algorithms tend to converge to a highly homogeneous population of solutions. In this context, we chose to study the contribution of Determinantal Point Processes (DPP) in parent selection. These probabilistic models allow for random sampling within the population of solutions, with diversity properties among the selected solutions.