Importance identification for fault trees based on possibilistic information measurements

Abstract

Subjective factors and nonlinear characteristics, inherent in the importance identification for a fault tree in the reliability and risk analysis, make it necessary for fuzzy (or possibilistic) approaches to accommodate the quantificational assessment of epistemic uncertainty in a practical problem when data and information are very limited. After investigating the intuitive interpretations, possibilistic information semantics, measure-theoretic terms and entropy-like models, a new axiomatic index of importance measure for fault trees is proposed based upon possibilistic information entropy, which adopts the possibilistic assumption in place of the probabilistic one. An example of the fault tree is provided along with the concordance analysis and other discussions. The more conservative numerical results of importance rankings that involve more choices could be viewed as “soft” fault identification under a certain expected value. Finally, possible extension to the evidence space and further research directions are discussed.