Generalized dynamic programming for multicriteria optimization (Q908857)

This paper is an interesting contribution to the applications of dynamic programming to vectorial problems decision. The authors show that the hypothesis of monotonicity which is essential but difficult to verify even for usual multicriteria preference functions in some finite vectorial optimization programs in order to obtain optimal solutions remains only a sufficient but not necessary condition for the optimality. Together with an original algorithm for a combination of a shortest path and reliability type problem, this methodology is also important for the study of the properties of sets of optimal solutions. On the other hand, the flexibility of this generalized dynamic programming method suggests a new way for the investigation of multicriteria optimization problems in infinite dimensional spaces, although it seems to be applicable only for finite multicriteria sequential vectorial decision problems.