Direct risk neutral stellarator coil design

Designing stellarator coils for fast particle confinement is a complex optimization problem with multiple local minima. The optimized coils can be very sensitive to manufacturing defects and misplacements, and stringent engineering tolerances are needed to ensure high performance of the stellarator. The goal of risk neutral coil optimization is to find stellarator coil sets that are less sensitive to these errors. Recent works have formulated this as a stochastic optimization problem seeking stellarators with quasisymmetry near the magnetic axis. In this work, we present and then compare a few techniques to solve these computationally intensive problems beyond the near-axis framework.