GRANULAR-INFORMATION-BASED RISK ANALYSIS IN UNCERTAIN SITUATIONS

Abstract

In the real life almost all of the decisions that we have to make incorporate uncertainty about the future events. Assessment of the uncertainty and, thus, the risk that is inherent in these decisions models can be critical. It is even truer if we are talking about the possibility of negative impact on the environment. It is very important to assess all the environmental risks in a project if there is any hazard to the environment. In this paper the possibility of using granular information is considered. The main advantage of the granular information is that it can be used to assess risks in situations when information about future events is incomplete and imprecise. Moreover, we can use natural language to describe the problem area, as granular information paradigm uses both fuzzy and probabilistic information. We propose to use entropy as the measure of uncertainty. However, the definition of entropy should be generalised, as values of probabilities, upon which the calculation of entropy is based on, are interval-valued. We propose several possibilities of generalizing the definition of entropy. Furthermore, we analyse these approaches to see whether the additivity feature holds for the generalized entropy.