Two-level orthogonal screening designs with 80, 96, and 112 runs, and up to 29 factors

Abstract

Due to recent advances in the development of laboratory equipment, large screening experiments can now be conducted to study the joint impact of up to a few dozen factors. While much is known about orthogonal designs involving 64 and 128 runs, there is a lack of literature on screening designs with intermediate run sizes. In this article, we therefore construct screening designs with 80, 96 and 112 runs which allow the main effects to be estimated independently from the two-factor interactions and limit the aliasing among the interactions. We motivate our work using a 14-factor tuberculosis inhibition experiment and compare our new designs with alternatives from the literature using simulations.