I am currently working on a scenario generation-portfolio optimization model. I am generating scenarios using different stochastic methods by looking back at 252 days of data and generating 1000 simulations of the next 21 days of returns. Hence, this results in an array that is (n_sim * n_days * n_assets). In order to compute the optimal portfolios most definitions only take a mean vector and covariance matrix from a 2-day array (n_days * n_assets). How can I use my simulations effectively? I am working in python. Continue reading...
