A comparison of frequency domain methodology and conventional factor screening methods

This paper presents a large sample comparison of two techniques in computer simulation factor screening: frequency domain methodology and the conventional two-level fractional factorial. The comparison is a cost versus power of the test comparison where a fixed budget is established and the resulting power of the test for each approach is analyzed and compared. In addition, the paper notes the similarities and differences of these two approaches. It also shows how run length, number of factors, number of runs, run start-up costs, factor variation, and system memory influence this comparison.