Practical Aspects of Algorithmic Design of Physical Experiments (from an Engineers Perspective)
Abstract: Due to operational or physical considerations, standard (i.e. canned) response surface and mixture designs often prove to be unsuitable for actual experimentation. In such cases an algorithmic design is required. I will explore various mathematical properties useful for evaluating alternative algorithmic designs. To assess “goodness of design” such evaluations must consider the model choice, specific optimality criteria (e.g. D, IV, etc), precision of estimation (fraction of design space), the number of runs (required precision), testing for lack of fit, and so forth. These issues are considered at practical level – keeping the actual experimenters in mind. This brings to the forefront such considerations as subject matter knowledge (first principles), factor choice and the feasibility of the experiment design.