
SCALARIZATION OF THE VECTOR PREFERENCES: OVERCOMING PRIMITIVIZATION
Author(s) -
Д. В. Трошин
Publication year - 2014
Publication title -
strategičeskie rešeniâ i risk-menedžment
Language(s) - English
Resource type - Journals
eISSN - 2618-9984
pISSN - 2618-947X
DOI - 10.17747/2078-8886-2013-3-88-94
Subject(s) - pairwise comparison , entropy (arrow of time) , scalar (mathematics) , basis (linear algebra) , computer science , substitution (logic) , econometrics , mathematics , machine learning , data mining , mathematical optimization , artificial intelligence , physics , geometry , quantum mechanics , programming language
It considers the drawbacks of the method for a scalar function preferences defined in a vector space on the basis of additive form and coefficients of importance. The advantages of the approaches based on the use of utility functions, replacement rates, pairwise comparison of indicators are discussed. Proposed ways of levelling shortcomings of the method of pairwise comparisons. As an alternative approach proposed target approach to the definition of the functions of preferences on the basis of the description of the physical effects of the system's functioning and use of criteria for substitution. Target approach considered on the example of the use of the entropy measures to assess the quality of surveillance. The hypothesis about the possibilities of cataloguing the main systems and development models scalarizаtion vector preferences using a targeted approach.