Design of experiments for project managers
This is the first article in a series about the design of experiments (DoE) concept and Taguchi methodology. Project managers are frequently called upon to make trial adjustments in project plans in order to reduce cost or completion time. Four approaches are available: trial and error, Monte Carlo simulation, sensitivity analysis, and design of experiments. The DoE approach is that measurements will be compared to draw general conclusions about system behavior. Each part in a manufacturing process, for example, is a variable, or factor. The system requirements for the application of experimental design are: the factors must be controllable; they must be adjustable to at least two values, each factor setting must be independent of the others; the output must be quantitatively measurable; and it must be possible to perform trials.