Comparison of different screening methods

Abstract

The aim of this paper is to illustrate sensitivity methods using experimental designs. These techniques are used to deal with simulation models or experimental programs involving a large number of controllable factors. In these cases, one of the objectives is the identification of the subset of factors that have substantial influence on the responses. In this paper we provide a review of different screening methods that are useful in eliminating negligible factors so that efforts may be concentrated upon just the important ones. These different methods are described and are applied on a simulation model in the petroleum industry involving 51 parameters. The comparison of the results is presented with the advantages and disadvantages for each method. The presentation is accessible to readers with an intermediate level of statistics.