An improved DEMATEL method for multigranular hesitant fuzzy linguistic environment

The decision-making trial and evaluation laboratory (DEMATEL) is widely recognized as an effective way to identify the key elements of a complex system and establish a causal relationship between the criteria. To extend the effective use of the DEMATEL method in multicriteria group decision making problems, we propose some improvements for traditional DEMATEL. The direct-influence matrix is a pairwise comparison matrix of the influence level between criteria, which is the basis of the DEMATEL method. Instead of utilizing crisp numbers to assess the level of direct-influence, each expert provides evaluation with hesitant fuzzy linguistic term sets (HFLTSs) based on an individual linguistic term set, that is, multigranular hesitant fuzzy linguistic term sets. To obtain a reasonable and satisfying collective DIM, we propose the expert weights model based on the similarity and entropy of HFLTSs and introduce the consensus reach process that is an iterative and dynamic process of modifying experts’ opinions more similar or closer to obtain the satisfying group solution. Considering the increasing complexity of the decision environment and the intricate relationship between the criteria, the Otsu method is applied to determine the threshold value of the total-influence matrix that is the sum of the direct influence and indirect influence and thus eliminate some unnecessary influence. Finally, the feasibility of the improved method is illustrated by an example of information warfare and analysis. Comparative results are presented to show the effectiveness and advantages of the proposed method.