Haralambos Evangelaras
Screening with Orthogonal Arrays
Screening designs are useful in situations where a large number of factors (q) is examined but only few (k) of them are expected to be important. It is desirable to know the full list of possible such designs in order to find the most efficient one that covers our experimental standards. Orthogonal arrays have traditionally been used as screening designs since they provide uncorrelated estimates of the main effects. We discuss the properties of a screening design and we describe an efficient algorithm we use for the identification of all possible orthogonal arrays with specific parameters. As an application, we outline their use in robust parameter design.
