The research of possibilities for fast calculation of median consensus rankings

Abstract

<p>We investigated the possibility of fast computation of collective expert estimates of the median type. Despite the widespread use of the Kemeny-Snell and Cook-Seiford medians for calculating the collective expert estimates, the possibilities to reduce the time required to calculate median consensus ranking through the application of the assignment problem and known algorithms for solving it, have been insufficiently investigated. In contrast to the most known methods, the method proposed in this paper is not approximated and retains the original median axiomatics of Kemeny. We investigated a possibility for calculating the Kemeny-Snell and Cook-Seiford medians using the assignment problem applying methods of computer experiment. We estimated the time required to calculate median rankings by four different algorithms for solving the assignment problem. It was established that the proposed method, at a moderate number of alternatives (n&lt;50), computes median rankings over the time close to a real time mode. It is also shown that, in contrast to other methods, the calculation of median rankings using the assignment problem does not depend on coherence degree of individual expert rankings. The results obtained are useful for the practical application of the examined procedure in network expert systems. In such systems, the computation time of a consensus ranking should be close to real time. In addition, in the network expertise systems, due to the random gathering of a team of experts, there is a possibility of the low level of coherence in individual rankings. For the examined procedure, that does not affect the duration of computation. This allows us to recommend the developed computational procedure for a fast search for median consensus rankings by Kemeny-Snell and Cook-Seiford for practical application in the systems of network collective expertise.</p>

Authors and Affiliations

Viktor Boltenkov, Varvara Kuvaieva, Oleg Galchonkov, Alesya Ishchenko

Keywords

Related Articles

Surface hardening and finishing of metallic products by hybrid laser­ultrasonic treatment

<p>Theoretical and experimental study of the possibilities of using laser heat treatment (LHT) combined with ultrasonic impact treatment (UIT) for surface hardening and finishing of metallic products was carried out. The...

Simulation of a gas and air flow exhausted by production equipment

<p>The paper provides the results of researching the influence of the gas and air flow at the exit from the pipeline of the technological channel on the performance of a gas and air power unit. It has been determined tha...

Studying and substantiation of the method for normalization of air­ionic regime at industrial premises at the ultrasonic ionization of air

<p>We report results of investigations on normalization of the air-ionic mode at premises with ultrasonic ionization of humidified air. We substantiated an increase in the concentration of negative air-ions with complex...

Development of the unified model for identification of composition of products from incineration, gasification, and slow pyrolysis

<p>This paper considers the processes of incineration, gasification, and slow pyrolysis. The common approach implies the use of individual models for the description of separate processes. When parameters acquire the val...

Modelling the loading of the nose-free cutting edges of face mill with a spiral-stepped arrangement of inserts

<p>A mathematical model of loading of the nose-free cutting edges of each insert of the face mill with a spiral-stepped cutting scheme was created. It was established that the size of the cut elements depends on the feed...

Download PDF file
  • EP ID EP528069
  • DOI 10.15587/1729-4061.2018.140686
  • Views 62
  • Downloads 0

How To Cite

Viktor Boltenkov, Varvara Kuvaieva, Oleg Galchonkov, Alesya Ishchenko (2018). The research of possibilities for fast calculation of median consensus rankings. Восточно-Европейский журнал передовых технологий, 4(4), 27-35. https://europub.co.uk./articles/-A-528069