Abstract:
|
In planning a fractional factorial experiment, prior knowledge may suggest that effects involving some factors are more important than others. Literature on how best to incorporate such information when assigning factors to the columns of a design has not received much attention. We propose the concept of individual clear effects (iCE), which is in the same spirit of iWLP of Li et al. (2015) and iGWLP of Li et al. (2018). We discuss how to utilize iCE to assign factors more effectively. Motivated by a real problem, we introduce the clear effects pattern to construct the maximized clear effects pattern (MCEP) designs. These designs are often different from commonly used minimum aberration designs, and many of them also maximize the number of clear two-factor interactions. We then extend the definition of iCE and MCEP designs by considering blocking schemes. Finally, we study some properties of these designs, which can be used to reduce the computational burden for design construction.
|