Title: Computing optimal designs of multifactor experiments
Authors: Lenka Filova - Comenius University in Bratislava (Slovakia) [presenting]
Radoslav Harman - Comenius University Bratislava (Slovakia)
Samuel Rosa - Comenius University in Bratislava (Slovakia)
Abstract: A method is described for computing efficient approximate experimental designs for models with multiple independent discrete factors. The algorithm that we propose can be used to solve problems with an immense number of combinations of factor levels, e.g., 5 factors, each with 1000 levels, and obtain an optimal design in several seconds. The proposed algorithm alternates between two key steps: 1) the construction of exploration sets composed of star-shaped components and separate, highly informative design points and 2) the application of a conventional method for computing optimal approximate designs on medium-sized design spaces. We illustrate the performance of the algorithm on several nonlinear statistical models used in practice.