Lower-Order Confounding Properties of Inverse Yates-Order Designs with Three Levels

It is important to consider the confounding information of lower-order component effects when choosing the optimal design in three-level regular designs. This paper studies a class of three-level inverse Yates-order designs D_q(n), where q and n are the numbers of independent columns and factors, respectively. The lower-order confounding information of designs D_q(n) are given according to the three cases: (i) q<n<3^q-1,n=2k,k\in N; (ii) q<n<3^q-1,n=2k+1,k\in N; (iii) 3^q-1\leq n<(3^q-1)/2. The above results are illustrated by examples.