Validating fuzzy multicriteria analysis using fuzzy clustering

Chung-Hsing Yeh^* and Yu-Liang Kuo

School of Business Systems, Monash University, Clayton, Victoria 3800, Australia

Received 30 March 2003; revised 17 July 2003; accepted 2 September 2003.

Abstract

  Fuzzy multicriteria analysis (MA) methods suitable for a given decision problem usually differ in their normalization and aggregation processes for handling the alternatives’ performance ratings and criteria weights. Due to their structural differences, these methods often produce inconsistent ranking results for the same fuzzy MA problem. This paper presents a validation procedure using fuzzy clustering for selecting among inconsistent ranking results produced by different fuzzy MA methods for a given problem. The procedure compares the ranking results obtained by fuzzy MA methods using different normalization processes and aggregation algorithms with the clustering results of the alternatives by fuzzy clustering. An empirical study of evaluating Asia-Pacific major international airports is conducted to demonstrate the effectiveness of the validation procedure.

Keywords: Multicriteria Analysis, Fuzzy Sets, Fuzzy Clustering, Validation

* Corresponding author. Tel.: +61-3-9905-5808; fax: +61-3-9905-5159; email: chunghsing.yeh@infotech.monash.edu.au