期刊:Encyclopedia of Statistics in Quality and Reliability日期:2007-12-14被引量:2
标识
DOI:10.1002/9780470061572.eqr007
摘要
Abstract If only a fraction of all the treatment combinations of a full factorial design are selected to be run, the resulting design is called a fractional factorial design . By carefully deciding which fraction of the treatment combinations to run, it is possible to construct fractional factorial designs with good properties. We examine these designs in detail for designs in which each factor is studied at two levels. We show how to construct these designs and how to find their properties. We discuss some uses of these designs as well as their advantages and disadvantages. An example is given of a design and its analysis. Extensions are considered to other situations, including three‐level designs, blocked designs, and split‐plot designs.