**
Computational finance. Portfolio optimization. Model Predictive Control. Multiobjective optimization.
Combining complex mathematical models that optimize investment portfolios with efficient and practical computational techniques is possible with the advent of technology. The aim is to build a computational framework capable of dealing with the actual peculiarities of the financial world while using available theoretical knowledge. Therefore, this work proposes a computational framework composed of different models that use Model Predictive Control (MPC) to optimize investment portfolios. An innovative strategy that combines MPC and multi-objective optimization and realistic financial market constraints, such as investment limits, automatic finance, cardinality, and transaction cost, is proposed. Performance criteria (objective functions) are the expected values of wealth, variance, and CVaR. We proposed multiperiod formulations for the objective functions and the genetic algorithm that performs the optimization since the multiperiod perspective is essential for the MPC strategy. With the experiments performed, it was possible to obtain various findings involving the prediction horizon, cardinality, risk, and return of the portfolio. Among those that can be highlighted, the prediction horizon benefits the portfolio optimization problem. It offers wealth value not obtained by the myopic strategy; portfolios with smaller cardinalities have greater risk because they have capital allocation in the risk-free asset. The five cardinality portfolios have the highest wealth values. In the out-of-sample analysis, the accumulated wealth of the proposed strategy surpassed Ibovespa and prestigious investment funds in 2020.