S u models of the maximum of several dependent random variables: development and computer application

Abstract An approach for modeling the distribution of the random variable ( U ) that is the maximum of several dependent random variables is described in this paper. Applications of the model to various problems associated with the Industrial Engineering profession are described and one — the accumulation of components in small-lot assembly systems — is studied in detail. Numerical tests investigated the accuracy of the approach and identify several fundamental characteristics of the accumulation process. Apparent advantages of the model are that it requires computer run time that is fraction of that required by Monte Carlo sampling (simulation) procedure and that it is readily amenable to parallel processing.