New Improved DEMATEL Method Based on Both Subjective Experience and Objective Data.

Yuan-Wei Du,Wen Zhou
DOI: https://doi.org/10.1016/j.engappai.2019.05.001
IF: 8
2019-01-01
Engineering Applications of Artificial Intelligence
Abstract:The decision-making trial and evaluation laboratory (DEMATEL) method is a hot issue in industrial engineering field for it can help determine critical factors in complex system. Although lots of efforts have been spent on improving the DEMATEL, they are just the extensions from the subjective perspective but lack of the objective perspective. This study focuses on providing a new improved DEMATEL method based on both subjective experience and objective data. In order to reasonably determine the initial direct-relation (IDR) matrix, the basic probability assignment (BPA) function is employed to extract expert experience and the Dempster's rule with Shafer's discounting is employed to make combination to derive the subjective IDR matrix. Then the path analysis is suggested to test each possible influence relation included in the subjective IDR matrix, and the objective IDR matrix consisting of path coefficients of any two factors is derived by training the sample data of factors. Following the principle of one-vote negation, the Dempster's rule is once again used to make combination for two kinds of IDR matrices, based on which an algorithm for the new improved DEMATEL is summarized to find the major factors in a complex system. Finally, numerical comparison and discussion are proposed to demonstrate the applicability and superiority of the prosed method.
What problem does this paper attempt to address?