A new algorithm for IFFD generation

Abstract

The authors describes the automated generation of optimal irregular fractional factorial designs (IFFDs), and for regular maximal subset designs (MSDs), for processes involving 2-level factors. They describe mechanisms on separating out confounded effects and new, efficient methods for IFFD generation. It is concluded that, although finding IFFD solutions by using a seed set of tests as a restriction shows some promise as a way to limit the search space, much work needs to be orchestrated toward solving a requirement set containing higher-order interactions.<>