On the use of MANOVA in the analysis of multiple-response simulation experiments
Linda Weiser Friedman · 1985
The purpose of this paper is to demonstrate the applicability of multivariate analysis of variance (MANOVA) to simulation analysis. The paper discusses an illustrative simulation experiment of an M/M/s queueing system, in which three system configurations are crossed with two queue disciplines to produce a 3 x 2 factorial design. Since there are several output variables of interest, namely average number of waiting demands, average time in system, and the proportion of demands forced to wait more than a specified amount of time, a multivariate statistical technique must be used. The necessity and advantage of using this multivariate tool, rather than treating the simulation experiment as several univariate-response experiments, is discussed.