SELECTION OF MULTI-CRITERIA ANALYSIS METHODS ON THE EXAMPLE OF THE PROBLEM OF RANKING

Abstract

This work is devoted to the selection and comparison of popular traditional methods of multi-
criteria decision making. The article presents an overview of the existing works of recent years
on the topic of their comparison, highlights the main criteria, as well as the most significant results.
Further, an example of the implementation of a DSS (decision support system) was considered
on the recommendation of such a method to the user, which includes a description of not only
the main methods, but also their modifications, highlighting an exhaustive taxonomy of multicriteria
analysis methods in general. For the selection of methods in this article, international
databases of scientific publications were used: Science Direct, Google Scholar and IEEE Xplore.
Certain search settings have been made to retrieve jobs that match the query. The next step describes the task of ranking alternatives to demonstrate the results of applying the selected methods.
As a method for distributing the weights of the criteria, the method of analysis of hierarchies
(AHP) was used. The calculation results are presented in tables and graphically. The evaluation
metric was considered to be the stability of the method to the number of alternatives and criteria,
as well as sensitivity to the weights of the criteria. At the current stage of the study, the following
methods were selected: TOPSIS, WASPAS, VIKOR, PROMETHEE and ELECTRE. As a result of
the study, optimal methods were determined (in terms of the ratio of computational complexity to
stability) for their further use in the development of DSS, the ELECTRE method was used as an
additional tool with a large number of alternatives to screen out the least attractive ones.
PROMETHEE showed high sensitivity to changes in weights and complexity of calculations, therefore
it was excluded from further development stages. VIKOR and TOPSIS showed the best stability
with the simplicity of calculations.