Split Plot Designs for Mixture Process Variable Experiment

Mika Alvionita Sitinjak¹, Utami Dyah Safitri2 and Erfiani3

^{1, 2} Departement of Statistics, IPB University, Bogor, Indonesia

(E-mail: mikaalvionitas@gmail.com, utamids@gmail.com)

³ Departement of Statistics, IPB University, Bogor, Indonesia

(E-mail: erfiani_ipb@yahoo.com)

ABSTRACT

The design mixture process variable has a problem if the process variable increases then number of experiment units increases too. So that, a design is needed  to control number of design, that design is split plot. D-optimal and G-Optimal criterion is used the main design criterion. This design criterion using point exchange algorithm. The D-Optimal design is design minimize the generalized variance of parameters estimates. The G-Optimal design is design minimizing the maximum prediction variance. This research focused how to optimization of split plot design for mixture process variable experiment. The present article were three component is a cookies recipe. The D-optimal design and G-optimal design have a same design point.