Selection Models Synthesis Based on Expert Estimates Extrapolation

DOI

10.2991/aebmr.k.200730.020How to use a DOI?

Keywords

generalized criterion, maximum likelihood method, approximation, Monte-Carlo method

Abstract

The analysis of large-scale business projects is non-dominant alternatives. However, the options under consideration may be too large, and the decision-maker may not be able to apply any mechanism for selecting the best option to this set. Most of the existing decision support procedures involve the entire available alternatives set in the comparison and evaluation process, so they are not suitable in this situation. The paper suggests an effective way to solve this problem – the expert assessments extrapolation method to develop an objective collective solution based on alternatives small training sample expert analysis. In the proposed method version, it is assumed that for any alternatives pair, experts are able to estimate the difference value in their utility. Thus, a difference-classification scale is introduced for alternatives, which makes it possible to more accurately assess the comparative alternatives value and make a more reasonable choice than when using an ordinal scale. This approach advantage also consists in the absence of any some alternatives superiority degrees priori numerical estimates over others, since any such assessment contains certain arbitrariness. The collective choice is based on obtaining generalized criterion parameters estimates using the maximum likelihood principle. In this case, calculating the likelihood function for m-alternatives sample requires determining the multiplicity m-1 of several integrals numerical values over complex geometry region. It is proposed for its calculation to use the Monte-Carlo method. To increase the stability to the maximizing the likelihood function method integration error, we propose numerically-analytical method for calculating target function first and second orders partial derivatives. Simple and visible examples demonstrate the proposed approach effectiveness.

Copyright

© 2020, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Y.V. Bugaev
AU  - S.V. Chikunov
AU  - B.E. Nikitin
AU  - M.N. Ivliev
PY  - 2020
DA  - 2020/08/01
TI  - Selection Models Synthesis Based on Expert Estimates Extrapolation
BT  - Proceedings of the Russian Conference on Digital Economy and Knowledge Management (RuDEcK 2020)
PB  - Atlantis Press
SP  - 108
EP  - 113
SN  - 2352-5428
UR  - https://doi.org/10.2991/aebmr.k.200730.020
DO  - 10.2991/aebmr.k.200730.020
ID  - Bugaev2020
ER  -