Discrete approximation in quantile problem of Portfolio selection (Q2752032)

The authors study the problem of optimal portfolio investment when expected return is considered as a criterion. They pay attention to paradoxical situations being caused by an unsuccessful choice of the objective in the form of expected return. In those cases expected return tends to infinity, but the probability of ruin tends to unity. To overcome this obstacle the probability (or quantile) criterion is used and a stochastic model with quantile criterion is presented. A confidence solution based on the confidence method and on discrete approximation of a confidence set is suggested. Two algorithms based on the confidence method and discrete approximation of the probability measure are proposed for the solution of the quantile problem. Several algorithms for solving the portfolio optimization problem are compared in an example.NEWLINENEWLINEFor the entire collection see [Zbl 0964.00055].