EXPERIMENTAL DESIGN FOR DEPENDENT DATA
S. G. Jagath Senarathne · Bulletin of the Australian Mathematical Society · 2020
This PhD focused on developing new methods to design experiments where dependent data are observed. Of primary consideration was Bayesian design, i.e. designs found based on undertaking a Bayesian analysis of the data. The generic design algorithms and the loss functions proposed in this study cater to a wide range of applications, including designing clinical trials and geostatistical experiments. These tools enable informed decisions to be made efficiently through maximizing the information gained from experiments while reducing costs.