PROPAGATION OF MEASUREMENT UNCERTAINTY EXPRESSED BY A POSSIBILITY DISTRIBUTION WITH COVERAGE-INTERVAL-BASED SEMANTICS

The main purpose of this paper is to present a possibility theory-based generalization of conventi o al interval propagation to coverage intervals. Indeed, the whole set of coverage intervals for all the probability l evels stacked on top of one another constitutes a possibi lity distribution. Thus by slight modifications of Zadeh ’s extension principle, we will prove that it is possi ble to compute the coverage intervals of an indirect measu rement from many other measurements (possibly dependent) b y a known non decreasing relationship.