An Approach to the Construction of Mixed Model Orthogonal Designs for Screening in Industrial Experimentation - The Method of Symmetric Constructions

摘要

Fractional factorials and orthogonal main effect plans are popularly used as design plans for screening experiments involving several factors, to identify the important factors influential in understanding a process. Such design plans, while being simple in concept and creation for symmetric experiments, pose interesting challenges in attempting a generic tractable solution for the case of asymmetrical experiments. It is the object of this paper to present a generic technique called the Method of Symmetric Constructions for the construction of mixed model design plans. This method is applicable for any factor combination and is not restricted to prime power factor levels alone. Also, several rules have been derived for factor replacement and collapsing that offer maximum flexibility in factor level selection to encourage the creation and use of mixed model orthogonal designs for industrial experimentation.