B1601
Title: MultiDOE: A multi-criteria design of experiments R package
Authors: Francesca Cucchi - University of Milan-Bicocca (Italy)
Andrea Melloncelli - Vanlog (Italy)
Francesco Sambo - Verizon Connect Research (Italy)
Kalliopi Mylona - King's College London (United Kingdom)
Matteo Borrotti - University of Milan-Bicocca (Italy) [presenting]
Abstract: Many real experiments involve some factors whose levels are more difficult to set than others, most times due to high cost and/or limited time. A possible solution is to design the experiment according to a multi-stratum structure, where restrictions on the complete randomization of the experiment limit the total number of hard-to-set factor level changes. Considering optimal experimental designs, multi-criteria optimization searches for the best trade-off between competing research objectives. The MultiDoE R package can be used to construct multi-stratum experimental designs (for any number of strata) that optimize up to six statistical criteria simultaneously. In the first place, the algorithm relies on a local search procedure to find a locally optimal experimental design. More precisely, it is an extension of the Coordinate-Exchange (CE) algorithm that allows both the search of experimental designs for any type of nested multi-stratum experiment and the optimization of multiple criteria simultaneously. In the second place, the final solution implements a Two-Phase Local Search framework, that can generate a good Pareto front approximation for the optimization problem under study. The package provides different ways to choose the final optimal experimental design among those belonging to the Pareto front.
