The aim of this thesis work is based on the search for a metaheuristic algorithm able to efficiently solve the K clusters with fixed cardinality problems, which are NP-hard combinatorial optimization problems. To do this, the mixed integer linear programming (MILP) formulations that were presented in the study conducted by Gonclaves & Lourenco, 2019 were used as reference models. Initially, the instances to test were generated randomly using the software MATLAB and the edge graphs present in the CEDRCIC’s library, considering the same characteristics as those present in the study. Subsequently the models of the paper were compared. Then an attempt was made to reduce the metaheuristic CPU execution times without straying too far from the optimal solution obtained by the MILP model and promising results have been achieved.