R-number Cognitive Map Method for Modeling Problems in Uncertainty and Risky Environment

In qualitative, non-numerical models, there are numerous challenges raised by the ambiguity, uncertainty, and risk within variables. The root of this ambiguity may lie in the variable itself or other related variables, or be due to expert opinions. Fuzzy cognitive maps allow better understanding of such problems, determining the cause-effect relationships between variables. When dealing with problems where the associated numerical data are unavailable, or the nature of the problem is qualitative, cognitive maps are constructed based on the statements of related experts. One of the problems with using the common cognitive maps model that it fails to consider uncertainty, risk and error within expert comments. This problem affects the quality and credibility of the models applied to complex issues. This paper proposes an R-cognitive maps approach, based on the distance-based automatic construction approach and R-numbers, in order to capture possible risks, uncertainty and ambiguity of expert opinions, with regard to variables and causality relationships. The proposed approach can function as a decision support tool for risk-based cognitive maps problems involving expert opinions, and is tested numerically with an engineering problem.