Optimal Designs For Heteroscedastic Mixture Models

This paper study two areas: (a) the optimal designs for the heteroscedastic modelare applied to the mixture experiments. According to its theory, we propose D-optimaldesigns algorithm for heteroscedastic mixture model and have theory proof of themethods;(b) under certain conditions, we construct D-optimal designs for additivemixture model using its boundary model.This article is divided into five chapters. The first chapter introduces the mixtureproblem, mixture designs of the common problems. The second chapter introducesstructuring D - optimal designs KWT theory and algorithms, the basic principles ofconstruction algorithm when the model is the homoscedasticity. In chapter 3, we studythe optimal designs for heteroscedastic mixture models and obtain general D-optimalalgorithm. In Chapter 4, we study D-optimal designs for additive mixture models. InChapter 5, we analysis real data examples using DE algorithm.