Time-series based multi-criteria large-scale group decision making with intuitionistic fuzzy information and application to multi-period battery supplier selection

Abstract

Battery is the core component of new energy vehicles. The choice of proper battery supplier is crucial for new energy vehicle companies since it directly affects the market share and core competitiveness. Some criteria of battery suppliers cannot be evaluated in the short period and need to be continuously evaluated at different periods (or time). Decision makers (DMs) may present some hesitancy when giving their evaluations. Thus, this paper formulates the battery supplier selection as a type of time-series based multi-criteria large-scale group decision making (LSGDM) with intuitionistic fuzzy information. A novel method is proposed to effectively resolve the battery supplier selection. Firstly, considering the importance of different periods, the comprehensive time weights are determined by combining the exponential decay model and the maximal entropy ordered weighted averaging (OWA) operator weights determining model. Secondly, DMs are divided into different clusters by using the proposed intuitionistic fuzzy Hamming distance-based fuzzy c-means (FCM) algorithm on each criterion. Combining the confidence levels of DMs and the corresponding clustering results, the DMs’ weights with respect to each criterion are determined. Then, the comprehensive criteria weights are obtained by integrating the intuitionistic fuzzy entropy and the constructed multi-objective programming model. Thirdly, the individual consensus level and group consensus degree at each period are defined, respectively. Then, a new identification approach is proposed to identify DMs to be adjusted. Subsequently, a total adjustment minimization model is built to adjust these DMs’ evaluation information. The optimal alternative is generated according to the collective global overall evaluations. Finally, a real example of batter supplier selection is demonstrated to verify the feasibility and superiority of the proposed method.