On Agent-Specific Fuzzy Entropy Functions

The concern of this article is to introduce versatile fuzzy and probabilistic-fuzzy entropy functions that could quantify the overall ambiguity of a system as perceived by the agent. The central principle behind the proposed functions is: besides the underlying ambiguity, an agent's inherent outlook toward ambiguity casts a great influence on the perceived entropy. The proposed functions also consider the tendency to exxaggerate/downplay the ambiguity. More specifically, the proposed entropy functions take as argument the perceived ambiguity value, rather than the actual membership grade. Besides, the relationship between the perceived membership grade and ambiguity is closely modeled to compute the overall entropy. The variants of the proposed functions for the probabilistic-fuzzy domain are also presented. A real case study is included to demonstrate the significance of the work.