Efficient uniform designs for mixture experiments in three and four components
Abstract
This paper gives four classes of uniform designs for mixture experiments and investigates the properties of these classes of designs using various mixture models. The four classes of designs are (i) mixture designs constructed using uniform designs based on cyclic latin squares, (ii) mixture designs constructed by projecting the uniform designs based on good lattice point method, (iii) mixture designs constructed by projecting the uniform designs based on cyclic latin squares and (iv) uniform mixture designs given by Wang and Fang(1990). For restricted exploration of mixtures, we used the transformation given by Saxena and Nigam(1977) with some modification and constructed designs for each of these four classes of designs. The design efficiencies of the most uniform designs of the four classes of designs are also computed and compared.
Publication Title
Progress in Applied Statistics Research
Recommended Citation
Aggarwal, M., & Singh, P. (2009). Efficient uniform designs for mixture experiments in three and four components. Progress in Applied Statistics Research, 11-26. Retrieved from https://digitalcommons.memphis.edu/facpubs/4566
