Seminar| Institute of Mathematical Sciences
Time: Thursday, April 23th, 2026,10:00-11:00
Location: IMS, RS408
Speaker: Guanzhou Chen, Nankai University
Abstract: This talk gives an overview of several projects that I have done in the past few years. A consistent and coherent theme in these projects is the idea of stratification in various experimental designs, including space-filling designs and order-of-addition designs. Orthogonal array-based designs are a popular class of space-filling designs for modern experimentation due to their attractive stratification properties. We first justify orthogonal array-based designs, particularly strong orthogonal arrays, under a broad class of space-filling criteria, which include commonly used distance-, orthogonality- and discrepancy-based measures. We then introduce linear allowable level permutations and a projection stratification enumerator to select strong orthogonal arrays for better space-filling properties. Next, we propose a new class of space-filling designs, called the generalized orthogonal array-based designs. The proposed designs achieve stratifications over s1×s2 and s2×s1 grids in two-dimensions, where s1 and s2 can be arbitrary positive integers. Finally, we extend the idea of stratification to order-of-addition experiments and show the proposed designs are not only economical in run size, but also robust to model uncertainty.