A MCMEIF-LT model for risk assessment based on linguistic terms and risk attitudes

摘要

Due to the limitation of assessors’ knowledge and the uncertainty of risks, risk assessment data reasonably are given in the form of linguistic terms, safety risk assessment in petrochemical industry is often a multi-criterion and multi-expert information fusion based on linguistic terms(MCMEIF-LT) problem. A novel model dealing with the MCMEIF-LT problem is presented in this paper. Firstly, the individual linguistic assessment distributions are fused to collective distributions and multiple criteria are fused to a comprehensive criterion. In the fusion process, the objective weights of assessment experts are calculated with using the credibilities of assessment data and the attitudes of decision makers are considered. Secondly, a Fuzzy Number Weighted Ordered Weighted Aggregation(FN-WOWA) operator which can transform a fuzzy number into a crisp value is proposed. In the FN-WOWA operator, the utility function can incorporate the assessors’ loss-based risk attitudes and the membership function can reflect the importance of the values in the integrated fuzzy number. Based on crisp values, a risk matrix is constructed. Finally, a real application is demonstrated to show the flexibility and practicality of the MCMEIF-LT model.