Title: Optimal designs for complex problems
Authors: Weng Kee Wong - UCLA (United States) [presenting]
Abstract: Algorithms are practical ways to find optimal experimental designs. Most published work in the statistical literature concern optimal design problems for a model with few factors or assume the model is additive when there are several factors. Nature-inspired metaheuristic algorithms are general and powerful optimization tools that seem to be under-utilized in statistical research. We describe a few of these algorithms, list their advantages over current methods and present optimal designs for a few complex biostatistical problems with real world applications.