Decision-Making Based on a Conditional Fuzzy Measure

Abstract

The application of a conditional logical formula of three-valued calculus in the decision-making system is considered. A conditional logical formula makes it possible to determine a conditional fuzzy measure on its basis, which is associated with the following positive aspects. First, there is no need for expert evaluation of the fuzzy measure of the truth of the conclusion for fuzzy premises, which reduces the degree of subjectivity and eliminates the need to ensure the completeness of statistical data, as well as the justification of completeness. Secondly, the proposed version of calculating conditional conclusions relatively simply allows for a multi-premise case and the ability to evaluate the importance of premises based on their priorities (in classical approaches like Mamdani, premises do not differ in their degree of importance for conclusions). Thirdly, there is no needto eva luate the degree of truth of the rules themselves for fuzzy conclusions. These advantages simplify practical use and ultimately improve the quality of decisions made, especially in the case of a large number of inputs (for exam ple, numbered in tens). An example of the practical use of the approach developed on the basis of a fuzzy conditional measure for making decisions about the correction of the learning process based on the testing results is given.