Using a genetic algorithm to generate Ds-optimal designs with bounded D-efficiencies for mixture experiments

Authors

  • Saranya Thongsook Department of Mathematics and Statistics, Faculty of Science and Technology, Thammasat University, Rangsit Center, Pathum Thani 12121, Thailand.
  • John J. Borkowski Department of Mathematical Science, Montana State University, Bozeman, MT, USA.
  • Kamon Budsaba Department of Mathematics and Statistics, Faculty of Science and Technology, Thammasat University, Rangsit Center, Pathum Thani 12121, Thailand.

Keywords:

exchange algorithm, genetic algorithm, optimal design, mixture experiment, subset optimality

Abstract

We propose and develop a genetic algorithm (GA) to generate Ds-optimal designs for constrained mixture regions when quadratic terms are of primary interest. Our method does not limit the selection of design points from a finite candidate, but allows points to be select points throughout a continuous region in selection and reproduction process of GA. Summaries of GA designs are reported for 3 and 4 components. Moreover, efficiencies are used to compare the performance of GA designs with designs generated from an exchange algorithm (EA) and to computed-generated designs (CGDs).

Downloads

How to Cite

Thongsook, S., Borkowski, J. J., & Budsaba, K. (2015). Using a genetic algorithm to generate Ds-optimal designs with bounded D-efficiencies for mixture experiments. Thailand Statistician, 12(2), 191–205. Retrieved from https://ph02.tci-thaijo.org/index.php/thaistat/article/view/34198

Issue

Vol. 12 No. 2 (2014): JULY

Section

Articles