A randomized version of the recently developed barycenter method for derivative-free optimization has desirable properties of a gradient search. We develop a complex version to avoid evaluations at high-gradient points. The method, applicable to non-smooth functions, is parallelizable in a natural way and shown to be robust under noisy measurements.
@article{arxiv.1801.10533,
title = {The Barycenter Method for Direct Optimization},
author = {Felipe M Pait},
journal= {arXiv preprint arXiv:1801.10533},
year = {2018}
}