A Construction of Cost-Efficient Designs with Guaranteed Repeated Measurements on Interaction Effects

• Activity Number: 156 - Statistical Aspects in Stochastic and Deterministic Simulation
• Type: Topic Contributed
• Date/Time: Monday, July 30, 2018 : 10:30 AM to 12:20 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Frederick Kin Hing Phoa*
• Companies: Academia Sinica
• Keywords: Orthogonal Array; Covering Array; Repeated Measurement; Factorial Designs

Abstract

This work introduced a useful class of cost-efficient designs for two-level multi-factor experiments. It provided guaranteed repeated measurements on all 2-tuples from any two factors and the number of repetitions was adjusted by the experimenters. Given the number of factors of interest, it utilized less resources than an orthogonal array while its repeated measurement provided a resistance towards outliers that a covering array failed to achieve. To bridge the wide spectrum between two extreme settings (orthogonal arrays and covering arrays) in terms of the number of repeated measures of tuples, we developed a systematic method to construct families of these designs, namely (supersaturated) repeated coverage design, with small run sizes under different number of factors and number of repetitions.