A problem of ranking interval objects in a multicriteria analysis of complex systems

Three models of ranking interval-defined objects based on pairwise relations of domination with respect to probability, mathematical expectation, and utility are considered. It is assumed that the likelihood of the location of the true values of these objects within intervals given for them can be described by triangular probability density functions with generally biased centers of distribution. The models proposed are designed for application in automated decision support systems in an interactive mode.