This work want to build an approach that can select the best algorithmic parameters of the optimization solver IBM ILOG CPLEX, for solving instances of a specific Hydro Unit Commitment problem, This goal is achieved by mixing Machine Learning and Mathematical optimization methods and algorithms.