Prof. Weng Kee Wong
University of California at Los Angeles
Fri, 19/07/2019 - 4:00pm
RC-4082, The Red Centre, UNSW
This talk reviews and discusses nature-inspired metaheuristic algorithms as general purpose optimization tools for solving problems in statistics. The approach works quite magically and frequently finds an optimal solution or a nearly optimal solution quickly. There is virtually no explicit assumption required for such methods to be applicable and the user only needs to input a few easy tuning parameters. We focus on one of the more popular algorithms, particle swarm optimization (PSO), and as an application, demonstrate its ability to find various types of optimal experimental designs for dose response studies, multiple-objective design
problems and other biomedical problems, including optimal designs for generalized linear models with several interacting factors and standardized maximin optimal designs, where effective algorithms to find them have remained stubbornly elusive until now.