Generation of a pseudo-random set with desired correlation and probability distribution

Abstract

The following competition was organized by the Business Section of the Netherlands Society for Statistics and Operations Research (VVS): maxi mize the output of a given simulation model by selecting the best combination of six inputs; only 32 runs are permitted. Twelve teams competed; these teams came from industry and academia. This paper is written by the winning team, ex plaining its design and analysis. That design pro ceeded in stages. First, a special design was used to estimate all main effects and two-factor inter actions (namely, Rechtschaffner's saturated de sign). Then quadratic effects were estimated by changing factors one at a time. Finally, the re sulting estimated second-order polynomial was used to estimate the optimal input combination. The paper presents a combination of design of ex periment techniques and common sense that may have more applications in solving real problems.