An efficient algorithm to find portfolio weights for the first degree stochastic dominance with maximum expected return

Unlike the mean-variance approach, the stochastic dominance approach is to form a portfolio that stochastically dominates a predetermined benchmark portfolio. This study is to search a set of portfolio weights for the first-order stochastic dominance with maximum expected return by managing the constraint set and the objective function separately. A nonlinear programming algorithm was developed and tested with promising results against Korean stock market data sets.