On the use of Multistage Scenario Trees for Optimal Portfolios in the BMV

摘要

This paper describes the application of a scenario tree to represent the discretization of the random return values and its probability in order to find an optimal investment strategy by maximizing expected wealth at the end of a multistage horizon, considering wealth, return, cash balance, withdrawals and upper bounds. The scenario tree was generated with simulation and randomized clustering. The performance of the linear optimization model on different scenario trees is illustrated using test examples obtained from the BMV (Mexican Bursaries Market). The resultant model is stochastic and the computational results indicate that the approach is promising for this sort of problems because easily allow the introduction of more specific and real conditions, as constraints in the model.