Type of mixture designs and optimization according to the D-optimal criterion using an exchange algorithm

Abstract

Most engineers and technicians seek to improve their products or production processes through experimentation. Unfortunately, the strategies commonly used to conduct these experiments are often costly, inefficient, and result in numerous difficult-to-exploit experiments. For these reasons, many engineers and technicians turn to experimental design, particularly Mixture Experimental Designs. In this dissertation, a detailed study of mixture designs was given, and an software application was developed by python programming language to handle mixture designs. This provides a good understanding of the contribution of computer science to experimental design and the assistance an experimenter can expect from it. Key Words : Experimental designs, mixture designs, optimality criteria, D-optimality criterion, exchange algorithms.