Hybrid correlated data in risk assessment

This paper presents a method for evaluating the risks in a situation where parameters in the calculation are expressed in the form of dependent fuzzy numbers and probability distributions. The procedure of risk estimation combines stochastic simulation with the execution of arithmetic operations on interactive fuzzy numbers. In order to define operations on such numbers, non-linear programming is used. Relations between the parameters presented in the form of fuzzy numbers and probability distributions are expressed by means of interval regression. The results of computations indicate that the relations between parameters have a significant impact on the ratios characterizing risk.