4. Conclusions. Outlook

In this article, we have shown that the shape of the experimental domain in formulation can be very simple (a triangular diagram, for example) or very complex (domain with constraints, mixtures of mixtures, very disparate ranges of variation of components, presence of manufacturing factors, etc.). At the same time, we have shown how empirical mathematical models can be developed in these different cases, to enable simulation or optimization of experimental responses. In addition to standard experiment matrices, we presented techniques for constructing problem-specific experiment matrices, often referred to as "à la carte" matrices. Finally, we have taken into account the existence of several experimental responses and shown the possibilities for finding the best compromise.

In the following articles ( and on the planning of experiments...