Outline

  • Abstract
  • Keywords
  • 1. Introduction
  • 2. a Bird's Eye View of Criteria Interaction Phenomenon
  • 3. Overview of Dematel and Anp Methods
  • 3.1. Network Relation Map of Anp
  • 3.1.1. Related Review
  • 3.2. Inner Dependency of Anp
  • 3.2.1. Related Review
  • 3.3. Cluster-Weighted Anp
  • 3.3.1. Related Review
  • 3.4. Danp
  • 3.4.1. Related Review
  • 4. Bibliometric Analysis
  • 5. Concluding Remarks
  • Appendix A. Overview of Anp Method
  • Appendix B. Overview of Dematel Method
  • References

رئوس مطالب

  • چکیده
  • کلیدواژه ها
  • 1.مقدمه
  • 2. نمای دید پرنده پدیده تعامل معیارها
  • 3. مرور کلی روش های دیمتل (DEMATEL) و ANP
  • 1.3 نقشه روابط شبکه ANP
  • 1.1.3 مرور مقالات مرتبط
  • 2.3 وابستگی درونی ANP
  • 1.2.3 بازبینی مقالات مرتبط
  • 3.3 ANP وزن دار خوشه
  • 1.3.3 مرور مطالعات مرتبط
  • 4.3 DANP
  • 1.4.3 مرور مقالات مرتبط
  • 4. تحلیل کتاب سنجی
  • 5. نکات نتیجه گیری
  • ضمیمه A: مرور کلی بر روش ANP
  • ضمیمه B: مروری بر روش دیمتل

Abstract

Majority of the Multiple-Attribute Decision Making (MADM) methods assume that the criteria are independent of each other, which is not a realistic assumption in many real world problems. Several forms of interactions among criteria might occur in real life situations so that more sophisticated/intelligent techniques are required to deal with particular needs of the problem under consideration. Unfortunately, criteria interaction concept is very little issued in the literature. It is still a very important and critical research subject for intelligent decision making within MADM. The present paper aims to put a step forward to fill this gap by depicting the general picture, which provides a classification of methods related to criteria interaction phenomenon, and discuss/review the Decision-Making Trial and Evaluation Laboratory (DEMATEL) and Analytical Network Process (ANP) hybridizations first time in the literature. DEMATEL and ANP hybridizations grab remarkable attention of decision analysis community in recent years and seem as one of the most promising approaches to handle criteria interactions in a MADM setting.

Keywords: - - - -

Conclusions

In this paper a comprehensive analysis and review of the DEMATEL and ANP hybrid methods are provided. Four techniques are identified and carefully investigated along with their practical applications. The DEMATEL and ANP hybrid techniques are mainly used for the following needs:

• Identification of criteria structure: The ANP assumes that the criteria structure is known a priori. Despite the strong mathematical underpinnings of the ANP, it does not provide any systematic methodology to generate criteria and clusters, and to identify the corresponding interdependencies. For that reason, DEMATEL method is used to grasp NRM of the problem in most of the studies. We observed that, studies making use of DEMATEL for structuring the problem constitute 43% of the papers. This is the most commonly used way of DEMATEL utilization.

• Pairwise comparisons and survey questions: Especially, eliciting priorities for inner dependencies entails some ambiguous questions to be answered. Pairwise comparisons are, in general, cognitively demanding. However, when it comes to inner dependency, questions might be nonsense. To overcome this difficulty, inner dependency-related priorities in the supermatrix are calculated by DEMATEL. Since DEMATEL method is based on the cause and effect relationships, constructing direct relation matrix is much more practical. We observed that 8% of the papers make use of DEMATEL for obtaining inner dependency regions of the supermatrix.

• Unequal importance of clusters: In traditional ANP method, clusters are assumed to be equally important. However, clusters might have different priorities. For that reason, influences among clus-ters are obtained via DEMATEL method. This type of usage provides two important advantages. Firstly, it overcomes the unrealistic assumption of equally important clusters. Secondly, influence degrees among clusters are used to weight the unweighted supermatrix, which leads to weighted supermatrix. All columns of the weighted supermatrix are sum to unity so that it can be raised to large powers to obtain steady state weights. We found that this type of hybridization accounts for the 11% of the papers.

• Practicality concerns: One of the critiques toward ANP is that it is too complex to be applied practically. Researchers realized that the DEMATEL method indeed improves the ANP in terms of many aspects; structuring criteria, overcoming difficult pairwise comparison questions of inner dependencies, and weighting the unweighted supermatrix by unequal influences. In one step forward, DANP method is proposed that brings together all the advantages of other hybrid techniques. In DANP, criteria structure, unweighted supermatrix, and weighted supermatrix are all obtained via DEMATEL method, which is much more practical than traditional ANP. We observed that DANP papers comprise 38% of papers, and we expect growing number of applications of it in the years to follow. We also observed that the majority of the papers are originated from Taiwan. We expect that these techniques will be applied by researchers from other parts of the world for broader application fields. As a future study, investigating other hybrid techniques of DEMATEL method with other MADM methods would be interesting. We also emphasize that criteria interaction phenomenon is a crucial research topic, and should be investigated in much more detail. It is our belief that research in this direction will bring very interesting insights into the field of decision analysis.

دانلود ترجمه تخصصی این مقاله دانلود رایگان فایل pdf انگلیسی