Solutions of portfolio optimization problems are often influenced by model misspecifications or by errors due to approximation, estimation and incomplete information. The obtained results, recommendations for the risk and portfolio manager should be then carefully analyzed.
We shall focus on output analysis and stress testing with respect to uncertainty or perturbations of input data for the Markowitz mean-variance model with a general polyhedral convex set of considered portfolios and we shall discuss its robust versions. Possible extensions to general mean-risk efficient portfolios will be delineated.