This post is related to issue #19567 (https://github.com/scipy/scipy/issues/19567). The here-described optimization method is called the "Barycenter Method". His name originates from calculating the barycenter respective to some Oracle map based on evaluation points. Each point evaluates to a scalar quantity, and the barycenter weight is the exponential of the scaled negative of this value. The higher the value, the lower the weight. It has a gradient behaviour, but it derivative-free. I would be glad to have this method incorporated to scipy, since I worked with on my Master thesis. [1] https://arxiv.org/abs/2102.10467 Currently, I have this python version: https://github.com/asmove/pybary . I would be very grateful for any feedback in this regard. Thanks for reading!