Dynamic optimization with asymetrical penalties

Abstract

In this paper we discuss about dynamic programming models with a quadratic objetive function. An extension is suggested to relax the hypothesis of symmetric penalties. The extended model allows for a more accurate modelling of preferences.