The consistency index in reciprocal matrices: Comparison of deterministic and statistical approaches
When checking the inconsistency level of a positive reciprocal matrix Saaty uses a deterministic criterion based on two parameters, a benchmark (the average), and a consistency level, usually 10%. Using results from a simulation experiment with 100,000 positive random reciprocal matrices of size varying from 3 to 15, we developed a probabilistic criterion and compare it to Saaty's index. We found that if a positive reciprocal matrix is consistent according to the deterministic criterion is also consistent according to the probabilistic criterion only if we accept a higher than usual probability of Type I error. Reducing this error implies that the benchmark must be a small percentile of the probability distribution of the consistency index.
Year of publication: |
2008
|
---|---|
Authors: | Vargas, Luis G. |
Published in: |
European Journal of Operational Research. - Elsevier, ISSN 0377-2217. - Vol. 191.2008, 2, p. 454-463
|
Publisher: |
Elsevier |
Saved in:
Online Resource
Saved in favorites
Similar items by person
-
Offshore outsourcing decision making: A policy-maker's perspective
Tjader, Youxu Cai, (2010)
-
Reply to "Remarks on the Analytic Hierarchy Process" by J. S. Dyer
Harker, Patrick T., (1990)
-
Firm-level outsourcing decision making: A balanced scorecard-based analytic network process model
Tjader, Youxu, (2014)
- More ...