Entropic Criterion for Decision Making under Total Uncertainty
Keywords:
Fuzzy Set, Membership Function, Fuzzy Entropy, Alternative Solutions, Linguistic Meaning, Measure Opportunities Alternatives, Structuring, Integrated Value, Coordinate of the Center of GravityAbstract
Purpose: Choosing the best solution from many possible alternatives always occurs under uncertainty. The available methods of
solving this problem are based on various hypotheses about decision-making, often giving contradictory results. This does not conform
with the methodology of stability theory according to which only a methodically invariant result of data processing agrees with the
reality. The aim of this work is developing a decision-making method which would work under uncertainty, without using hypotheses
about the decision-making situation and in accordance with the stability theory methodology. Results: A multi-criterial decisionmaking
problem has been formulated for the case of total uncertainty, when the alternatives are structured on the base of fuzzy entropy.
The essence of the method is that the criteria compliance evaluations are represented as fuzzy numbers or as linguistic statements
formalized by fuzzy sets. This method, unlilke others, does not need hypotheses about possible decision-making situations and fits
the stability theory methodology, as fuzzy entropy calculation by different methods does not lead to contradictory results. The use of
fuzzy entropy as a criterion for structuring alternatives under total uncertainty has been demonstrated. A fuzzy entropy calculation
algorithm has been developed, with the criteria compliance evaluations presented in linguistic form. Examples are given in which,
unlike the other available methods, the use of fuzzy entropy gives unambiguous guidelines for the best decision-making. Practical
relevance: The proposed method of structuring alternative solutions under total uncertainty on the base of fuzzy entropy can increase
the validity of the decisions you make by providing the result invariance in regard to the original data processing method.