A fuzzy approach to experimental design to optimize one or multiple criteria

In this article, we present a part of our work on active learning of fuzzy rule bases.We propose to accelerate the optimization of one or more criteria thanks to an experimental plan built step by step, accompanied by a Sugeno model capable of predicting this criterion. Such an approach is useful, for example, in materials science and can help researchers converge more quickly on a material with desired usage properties, while maintaining interaction with them.