DESIGN COMPUTING

• An irregular shaped experimental region: In this situation, the experimental region is under constraint such as the one for designs of mixture.
• A nonstandard model: There are situations where experimenters due to their knowledge of the studied process might use models other than the commonly used ones.
• Unusual sample size requirements: Instead of adjusting the experimental material to the standard designs, experimenters could use a more flexible CGD.
• Augmenting a design: There are situations in which experimenters have to augment an existing design with additional design points or additional factors. Augmenting a design can easily be done with the help of a computer.
• Blocking a design: Most experimental material in agriculture and industry are heterogeneous in nature and therefore its statistical analysis benefits from blocking. Like augmenting a design, blocking a design can easily be done with the help of a computer.
• CGDs are better than standard designs: Most partially balanced incomplete block designs (PBIBDs) are inferior to the corresponding CGDs in terms of the goodness of the design. The same thing can be said about balanced fractional factorial designs. Recently, we have shown that all seven PBIBD-based Box-Behnken designs are not as good as the corresponding CGDs in terms of rotatability as well as D- and G-efficiencies.

Recognizing the importance of CGDs in research, Design Computing is set up to:

• Promote a new philosophy of designing experiments, i.e. Design for the experiment, do not experiment for the design;
• Promote the use of CGDs in industrial settings where it is difficult to adjust the experimental materials to published plans and in remote agricultural research stations where it is easier to get access to a PC than to various catalogues and professional journals to find a usable plan;
• Foster research and research cooperation in CGDs.

©2009 Design Computing