Analysis of the Criteria for Optimal Experimental Designs

摘要

Optimal experimental designs provide solutions to many situations in which standard experimental designrntechniques cannot be applied. Additionally, the use of optimal designs has grown substantially due to the increasingrnavailability of software for this purpose. Optimal designs are useful because they provide an economic and efficientrnmethod for selecting design points, and the theory of optimal designs refers to developing experimental designs thatrnare optimum with respect to some designated design criterion. This method of experimental design is powerfulrnbecause it leads to algorithms for constructing designs, which can be used in many different applications, yet arncomprehensive analysis has never been made in the literature. In this paper, we analyze these criteria by identifyingrnconceptual connections and by comparing mathematical structures. We then discuss potential application areas forrnoptimal experimental designs and draw observations for practical implementations.