EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

1740. Comparing Interval Priority Estimation Methods by Estimation Accuracy of Solution Set in Terms of Ranking Alternatives

Invited abstract in session MA-44: Preference Learning 1, stream Multiple Criteria Decision Analysis.

Monday, 8:30-10:00
Room: 20 (building: 324)

Authors (first author is the speaker)

1. Masahiro Inuiguchi
Department of Systems Innovation, Graduate School of Engineering Science, Osaka University
2. Akiko Hayashi
Graduate School of Engineering Science, Osaka University
3. Shigeaki Innan
Graduate School of Engineering Science, Osaka University

Abstract

The analytic hierarchy process (AHP) is a useful method for multiple criteria decision making. It provides a priority weight estimation method from a pairwise comparison matrix (PCM). A PCM evaluated by a decision maker (DM) is frequently inconsistent. The inconsistency is assumed to come from the evaluation errors. The priority weights have been estimated by minimizing the errors. However, human perception is vague. Assuming that the inconsistency comes from the vagueness, estimating interval propriety weights instead of crisp priority weights from a PCM was proposed. Many interval priority weight estimation methods have been proposed because the original interval priority weights estimation method tends to estimate insufficiently wide interval priority weights. Numerical experiments showed that ranking alternatives by interval priority weights estimated by several methods performs better than the crisp priority weights estimated by the conventional methods when we assume the DM’s vague evaluation. It is shown that the solution to the interval priority weight estimation problem is frequently non-unique although representative solutions have been used in the numerical experiments. In this presentation, we show first that the set of all rankings of alternatives under the estimated interval priority weights can be obtained easily. Using the set of all rankings in the numerical experiments, we confirm the advantages of several interval priority weight estimation methods.

Keywords

Status: accepted

Back to the list of papers