ISSN 2079-3537      

 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                                                                                             

Scientific Visualization, 2019, volume 11, number 4, pages 66 - 80, DOI: 10.26583/sv.11.4.06

Visualization Algorithms for Multi-criteria Alternatives Filtering

Authors: A.A.  Zakharova1,,, D.A.  Korostelyov2,B, O.N.  Fedonin3,B

A Keldysh Institute of Applied Mathematics Russian Academy of Sciences

B Bryansk State Technical University

1 ORCID: 0000-0003-4221-7710, zaa@tu-bryansk.ru

2 ORCID: 0000-0002-0853-7940, nigm85@mail.ru

3 ORCID: 0000-0002-3511- 699X, rector@tu-bryansk.ru

 

Abstract

The article considers a decision-making problem for multi-criteria problems characterized by a large number of criteria and alternatives. As one of stages in making decisions in such situations, it is proposed to use a filtering of alternatives based on their visual images. For these purposes, the paper describes main aspects of this approach, as well as a statement of the algorithms for constructing the corresponding visual images based on radar and radial diagrams. A description of data preparation for operation of visualization algorithms for multi-criteria alternatives with an aim of their filtering is given. The paper describes software AlternativesVisualizer that implements the proposed algorithms, as well as providing interaction with an expert for visual filtering of multi-criteria alternatives. Additionally, the capabilities of the developed software are described. They include filtering alternatives based on threshold values, as well as the possibility of conducting a series of experiments in order to obtain the union or the intersection of filtered sets of alternatives. An experimental study of the developed algorithms and approaches effectiveness based on a synthetic test and an experiment on visual filtering of real alternatives describing the accuracy of calculating the inviscid flow around a cone using several OpenFoam solvers (rhoCentralFoam, pisoCentralFoam, sonicFoam, rhoPimpleFoam, QGDFoam) is presented. Each solver is characterized by 288 criteria, and with the help of the developed algorithms during visual filtering, the clear predominance of two solvers over the others is established.

 

Keywords: choice set visualization, alternative visual image, visual filtering, multi-criteria alternatives, radial diagrams, radar charts.