Abstract

The aim of this paper is to determine a beta distribution of the first type, when a group of experts coincide in the minimum and the maximum values of a random variable (times, in the case of tasks; cash-flows, in the case of investments, etc.), but they disagree with the more likely values. For this purpose, after studying the main parameters of the PERT methodology, we define the bet value (which we identify with the mean) and we deduce its expression by means of the confidence that the expert has in the modal value given by himself. A variation of this approach rises when several experts give their opinion on the modal values and we define the distrust degree as the variance of the values contributed by themselves. This procedure is an alternative method to the fuzzy approach.

Key-words: Beta distribution, expert, PERT, bet value, confidence, distrust, fuzzy.

You must be a member to download the full article.