Development of group decision making model under fuzzy environment
Multi criteria group decision making (MCGDM) methods are broadly used in the real-world decision circumstances for homogeneous groups. Some decision-makers’viewpoints at times are more important or reliable than others, or they may differ in terms of the decision-maker experience, education, experti...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2011
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/41629/1/FK%202011%20122%20ir.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Multi criteria group decision making (MCGDM) methods are broadly used in the real-world decision circumstances for homogeneous groups. Some decision-makers’viewpoints at times are more important or reliable than others, or they may differ in terms of the decision-maker experience, education, expertise and other aspects. Thus,
a heterogeneous group of decision makers with dissimilar members has to be composed in MCGDM. Multi-dimensional personnel evaluation is one of the most critical decisions to make in order to achieve the organization goals. In many
situations, raters may decide on the basis of imprecise information coming from a variety of sources about ratee with respect to criteria. In fact, some criteria are
completely quantifiable, some partially quantifiable, and others completely subjective; moreover crisp data is inappropriate to model real-world circumstances.
Linguistic labels or fuzzy preferences are therefore, used to deal with uncertain and inaccurate factors involved and seem more reliable in complex group decision situations.
In this research, heterogeneous group decision making models under fuzzy environment for multi-dimensional personnel evaluation were proposed to compensate the differences of decision makers’ knowledge such as: education,expertise, experience and other aspects. A new fuzzy group decision making method was developed under the linguistic framework for heterogeneous group decision
making that aims at a desired consensus. The method allocates different weights for each decision maker using linguistic terms to express their fuzzy preferences for
alternative solutions and for individual judgments. Besides, the classical ordinal approach method under a linguistic framework is developed for heterogeneous group
decision making, which allows group members to express their fuzzy preferences in linguistic terms for alternative selection and for individual judgments. Furthermore, a
fuzzy extension of technique for order preference by similarity to ideal solution (TOPSIS) method under fuzzy environment was proposed. The method covers heterogeneous group decision making by considering the decision makers’ viewpoint weights. In order to solve the problem of discrepancy between decision making methods’ results, a new optimization method was developed, to aggregate the results’of different decision making models.
The four proposed methods were used in a case study. Proposed methods focused on the implementation of fuzzy logic approach in the personnel evaluation system.
Furthemore, personnel were evaluated from different points of view (supervisors,colleagues, inferiors and employee him/herself). A fuzzy Delphi method and linguistic terms represented by the fuzzy numbers were developed to elicit qualitative and quantitative criteria and assess criteria weights and relative importance of evaluation group’s viewpoints. Then, the proposed methods’ results were compared to already established methods. The study identified that the results of the proposed methods are closely related to other methods and the selections made by the proposed methods approximately are identical with the other already established methods. The Spearman’s rank correlation coefficient shows highly consistent rankings obtained by the methods. No significant difference in the ranking of the proposed methods and the other established methods was observed. The results of the problems solution based on the aggregated proposed model show that the aggregated model achieved the highest value in the Spearman’s rank correlation compared to the average method and Copeland function. Furthermore, the high Spearman’s rank correlation coefficient between the rankings supports the consistency of the results and similarity of applicability of the methods. |
|---|