Neutral atom quantum computers require atom sorting to be able to achieve defect-free arrays. However, preparing large arrays becomes difficult due to experimental sources of noise like vacuum-limited lifetime and heating from tweezer transfers. Here we work towards building noise-robust rearrangement algorithms through developing the atommovr software.
Through Monte Carlo simulations, we extract lower bounds for optimal time scaling, benchmark heuristic algorithms under noisy experimental conditions, and develop a novel dual-species algorithm, InsideOut, capable of preparing arbitrary target configurations without getting stuck.
See the full details in our paper, published in Quantum: https://quantum-journal.org/papers/q-2026-07-29-2177/
atommovr is designed to be built upon by the community! Up-to-date information concerning new features and opportunities for contribution can be found on the project's GitHub page.