Title: Consensus-based optimization
Author: Claudia Totzeck (University of Wuppertal)

We discuss a gradient-free particle method for minimizing non-convex objective functions that may live in high dimensions. Consensus-based optimization is formulated as SDE system and models the discussion of particles with the purpose of finding a consensus very close to the minimizer of the objective. Passing to the many particle limit allows for a PDE formulation that can be exploited to establish a sound mathematical convergence theory. Theoretical results are confirmed by numerical results.