A Novel Algorithm with Orthogonal Arrays for the Global Optimization of Design of Experiments

In this paper, we develop a novelal gorithm that determines the overall best parameter setting in the design of experiments. The algorithm begins with successive orthogonal array experiments and ends with a full factorial experiment. The setup for the next orthogonal array experiment is obtained from the previous experiment by either fixing a factor at a given level or by reducing the number of levels considered for all currently non-fixed factors. We illustrate this method using a light-gauge steel wall sound isolation system with four parameters, each with three levels. In a previous study, the full factorial of 81 experiments was conducted, and the optimal parameter settings were determined. With the proposed algorithm, we found the same result using 15 experiments. As a further comparison, we obtained the optimal settings using a traditional Taguchi method and found that they correspond to the 23rd experiment out of the 81 experiments when sorted by the objective (or quality) function. We conclude that the proposed algorithm can provide an accurate, fast, and economic tool for the global optimization of design of experiments.