Making Experimental Designs Robust Against Time Trend Nam-Ky Nguyen

In numerous experimental settings, engineers and scientists might be obliged to run their experiments in sequence and so the run order of the experimental units must be taken into account. This paper describes an approach to construct experimental designs such that the main effects, 2-factor interactions and quadratic effects are orthogonal or near-orthogonal to the linear and quadratic trends. Some constructed designs will then be compared with those of John (1990). Some trend-free Box-Behnken designs will also be given