There are typically two types of consistency of fuzzy preference relations (FPR), namely additive and multiplicative consistency. They are defined based on the assumption that decision makers are rational and can provide strictly consistent FPRs. To take into consideration the bounded rationality of decision makers, the current study relaxes this assumption and proposes a new measure called triangular bounded consistency for judging the consistency of FPRs. To define triangular bounded consistency, a directed triangle is used to represent three FPRs among any three alternatives, with each directed edge representing an FPR. The condition of restricted max–max transitivity (RMMT) in the directed triangle is quantitatively examined. Under the assumption that the bounded rationality of a decision maker is characterized by their historical FPRs, which are represented by directed triangles that satisfy RMMT, triangular bounded consistency is determined using the historical FPRs. We then illustrate how triangular bounded consistency can be used to verify the consistency of FPRs that are newly provided by decision makers and how to estimate some missing FPRs that are not provided by decision makers. Finally, to demonstrate the application of triangular bounded consistency of FPRs in multi-attribute decision analysis, we investigate a problem that involves selecting areas to market products for a company.
- Triangular bounded consistency
- Fuzzy preference relation
- Restricted max–max transitivity
- Bounded rationality
- Multi-attribute decision analysis