RSOME: Robust Stochastic Optimization Made Easy
Mean-Variance Portfolio
In this example, we consider a portfolio optimization problem discussed in Ben-Tal and Nemirovski (1999). Suppose there are n = 150 n=150 n=150 stocks, and each stock i i i has the mean return to be p i p_i pi and the standard deviation to be σ i \sigma_i σi. Let x i x_i xi be the fraction of the total wealth invested in stock i i i, a classic approach is to formulate the problem as a quadratic program, where a mean-variance objective function is maximized:
max ∑ i = 1 n p i x i − ϕ ∑ i = 1 n σ i 2 x i 2 s.t. ∑ i = 1 n x i = 1 x i