Sankhya: The Indian Journal of Statistics

1996, Volume 58, Series B, Pt. 2, pp. 231--239

RESTRICTED REGION SIMPLEX DESIGNS FOR MIXTURE EXPERIMENTS IN THE PRESENCE OF PROCESS VARIABLES

By

M. S. RAMCHANDRA MURTHY  and P. LAKSHMI MANGA, Osmania University

SUMMARY.  In the study of experiments with mixtures, the component proportions xi(i = 1, 2, ...q) satisfy ${x _ i}\leq 0$, $(\sum_{i}^{q} x_{i}=1$. The factor space is a (q-1) dimensional simplex. When the q component proportions are constrained by $m_{i} \leq x_{i} \leq M_{i]}(i = 1, 2, ..., q) the region of exploration is a restricted region of the simplex are denoted by RRSD. RRSD in the presence of process variables are denoted by M X P RRSD. In this paper, two M X P RRSDs are constructed.

AMS (1980) subject classification.  62K, 62K99.

Key words and phrases. Extreme vertices design, mixture variable, process variable, transformation.

Full paper (PDF)

This article in Mathematical Reviews.