Multicriteria optimization of a single-cell oil production

The search of an experimental design based on D-optimality criterion to improve a bioprocess is proposed. The chosen model is a second-degree polynomial one selected for its satisfying precision leading to an accurate representation of non-linear systems with a low number of parameters. The optimization is carried out thanks to a real encoding diploid genetic algorithm developed within our laboratory. This study shows the setting of a multiobjective optimization procedure for bioprocesses having, in all cases, several criteria to optimize. A new approach by Pareto's domain determination and ranking by Decision Engineering tool is presented.