A theory of portfolio revision: robustness and truncation problems

A set of decision criteria, different from bayesian methods is developed here, which is useful for making optimal portfolio decisions in a mean variance framework, These criteria which emphasize robustness and sensitivity to outliers are applied to revise a given size portfolio by including new securities or excluding old ones. A set of theorems is utilized to characterize situations when revisions are unnecessary due to the original portfolio being robust in some sense