We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Analysis of Errors in Priority Vector Estimation and Their Relationship with the Correctness of the Final Ranking of Decision Alternatives.
- Authors
Grzybowski, Andrzej Z.; Starczewski, Tomasz
- Abstract
Ranking creation is the main purpose of the multiple-criteria decision analysis (MCDA). In practice, it is typically achieved by estimation of the priority weights that reflect the importance of each of the available alternatives - this process is called prioritization. One of the most popular MCDA methodologies is the Analytic Hierarchy Process (AHP). All priority-weights-estimation technics that are used under the AHP scheme are based on the so-called pairwise comparison matrix (PCM). The PCM elements represent the decision maker's judgments about the priority-weights-ratios. They are typically expressed in values from a predefined set of numbers that is called a scale. Because of human brain limitations, it is natural that these judgments are usually erroneous, and consequently, the estimates of the priority weights are erroneous as well. It is well understood, however, that serious errors in judgments make the information contained in the PCM worthless. Thus the decision makers need a procedure that enables them to accept a given PCM or reject it as a useless one. This paper is devoted to the simulation analysis of the prioritization errors and their relationship with the correctness of the final ranking of decision-alternatives. Our simulation experiments reveal some interesting facts about the impact of the adopted scale on the priority-weights-estimation errors and allow us to formulate important remarks about the PCM acceptance procedure.
- Subjects
ERROR analysis in mathematics; ANALYTIC hierarchy process; DECISION making; COMPUTER software correctness; PULSE-code modulation
- Publication
IAENG International Journal of Applied Mathematics, 2019, Vol 49, Issue 3, p381
- ISSN
1992-9978
- Publication type
Article