I would like to evaluate how a model of 18 parameters (0 to 1) would fit some experimental data. Minimization is also constrained by very complicated relations linking the 1
