OPTIMAL FRACTIONAL FACTORIAL PLANS WITH BLOCKS

Blocking is one of the major tools for increasing the sensitivity of an experiment. In the context of fractional factorial designs too, blocking of the runs in the design often becomes necessary. A standard technique of blocking a fractional factorial plan is to use one of the factors as an indicator of the blocks. This method results in orthogonal blocking in the sense that the information matrix for factorial effects under a model with block effects remains the same as that under a model without block effects. However, orthogonal blocking has its limitations. In this talk, optimal block designs for fractional factorial plans are discussed when the blocking is not necessarily orthogonal. Examples are given.