Determination of Pareto-Optimal Solutions to the Problem of Multicriteria Selection

The problem of multi-criteria selection with the construction of a Pareto-optimal subset of solutions is considered. The presence of a set of criteria requires the construction of procedures that ensure the determination of adequate evaluations of alternative solutions. Well-known approaches to justifying solutions to selection problems suggest using a comparison of selection options based on selected criteria and the criteria themselves. Most of these approaches are on the use of expert assessments to determine the merits of individual options based, but their application can be appropriate only in the absence of reliable data on the actual characteristics of the alternatives considered. These approaches often rely on assumptions that may not hold. This is the possibility of comparisons, the use of fixed scales to determine the level of preferences, assumptions about the linear dependence of evaluations according to criteria in different ranges of values, the use of unreliable samples, etc. The article proposes an approach to the formation of solutions to the choice problem based on the construction of a criterion that determines the assessment of the approximation of solution candidates to the virtual example. Mathematical models were of the dependence of evaluations by criteria on the characteristics of the objects of comparison built. The results of determining the architecture of the virtual data processing system from cloud providers from the positions of a number of criteria are given. The results obtained indicate the feasibility of applying the proposed approach to solve a wide range of selection problems.