-
Experiments with Replicate Runs at the Cube and Star Points of Inscribed Central Composite Designs
Julius C. Nwanya1*, Bolarinwa O. Ajala2, Chinyere G. Ihuoma1
1. Department of Statistics, Imo State University, Owerri.
2. Department of Mathematical Science, Lagos State University of Science and Technology.
Abstract: Replication has been a fundamental aspect in any design of experiment. This study looks at the impact replication have on inscribed central composite design (ICCD) when the cube and star points are replicated. The D-, A-,and G-optimality criteria were applied to find the optimal ICCD for experiments when two and three replicate runs are made at the cube and star points of the design at several levels of factor (k = 2 – 7). Results showed that Inscribed central composite design with replicated cube points (ICCD-RF) reduces the A-optimal values of the design in all the factors of considered in the study whereas at lower factors of k (k = 2 and 3), it also reduces the D-optimal values of the design and also increases the G-optimal values for most of the factors considered in the study. Also replication at the star points of Inscribed central composite design (ICCD-RA) increases the A- and D-optimal values of the design in all the factors considered in the study while for G-optimality criterion, replication at the star points increases the criterion values but not in all the factor levels.
Keywords: Replication, Inscribe Central Composite Design.
Pages: 63 – 71 | Full PDF Paper