The Application of Distance Measure on Group Decision-Making Problems and Similarity Measurement

• Main Authors: Wan-Lin Chen, 陳宛琳
• Other Authors: Liang-Hsuan Chen
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/82619390281864021434

碩士 === 國立成功大學 === 工業與資訊管理學系碩博士班 === 94 === 　　The distance measures are usually used as the similarity indicators in the pattern recognition. The objective of this research is to investigate the feasibility of eight existing distance measures in the aggregation of group fuzzy decisions and in the similarity measurement of two fuzzy numbers. In the first part of this research, we apply the concepts of α-cuts in each distance measure, and determine an optimal group fuzzy consensus based on mathematical theorems, representing the aggregation of individual fuzzy opinions. The ratio of the left distance to the right distance between the group fuzzy consensus and all individual opinions is developed to evaluate the quality of group fuzzy consensus. Using two datasets from the existing study and considering both weighted and original individual fuzzy opinions, the quality of group fuzzy consensus from eight distance measures is compared based on the proposed ratio. The best distance measure is adopted to compare with the existing aggregation methods of group fuzzy decisions. 　　In the second part, we also apply the concepts of α-cut to distance measure in measuring the degree of similarity between two generalized fuzzy numbers. The proposed method is demonstrated by comparing with the existing similarity measures using twenty-one sets of paired fuzzy numbers. The results show that the proposed similarity measure is better than the existing methods.