Title: A proposed voting scheme to reduce the sensitivity of the Markov method

Authors: Baback Vaziri; Yuehwern Yih; Tom Morin

Addresses: School of Industrial Engineering, Purdue University, West Lafayette, IN 47907, USA ' School of Industrial Engineering, Purdue University, West Lafayette, IN 47907, USA ' School of Industrial Engineering, Purdue University, West Lafayette, IN 47907, USA

Abstract: The Markov method is one of many successful ranking methods that uses Markov chains to obtain its ratings and rankings of alternatives. It has been shown, however, that the method is sensitive to upsets, particularly in the tail of its ranking. The method also exhibits faulty behaviour when it has a periodic Markov chain. This study proposes a modification to the voting scheme of the Markov method that will alleviate the sensitivity to upsets and remove the issue of periodicity in the Markov chain. To examine the sensitivity, we first provide an example and see how both voting schemes react to an upset. Next, we generalise both voting schemes and examine a ratio of rating increments to understand why the tailing effect occurs, and how we can subside its effect.

Keywords: ranking methods; Markov method; sensitivity; ranking vectors.

DOI: 10.1504/IJOR.2018.091200

International Journal of Operational Research, 2018 Vol.32 No.1, pp.24 - 40

Received: 17 Apr 2015
Accepted: 12 Jun 2015

Published online: 16 Apr 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article