Minimax approach for using the qualitative preferences in the multicriteria evaluation

—Decision support systems in analytical systems based on the use of big data involve the formation of integral assessments of population objects using all parameters or some subset of them. The article discusses the problem of obtaining a multi-objective (multi-parameter) assessment of objects and an approach that involves the use of importance weights in the presence of high-quality and, possibly, incomplete information about the relative importance of certain criteria. The fundamental principle of various quantitative assessments of the mutual preference of private particular criteria for various objects in the population is considered while maintaining the system of preferences of the entire set of objects. The approach used assumes that the decision maker formulates qualitative information about the relative preference of certain criteria in the form of a not necessarily complete preference graph. For each object, weighting coefficients are calculated automatically according to the principle of a guaranteed result by solving an optimization problem using generalized logical criteria of maximum risk and maximum caution. For special cases of preference systems, analytical relationships and algorithms for calculating weight coefficients are given. This technique ensures the use of additional qualitative information about the preferences of certain criteria, obtaining numerical values of significance weighting coefficients and solving the problem of multicriteria assessment based on the principle of a guaranteed result.