atommovr: An open-source simulation framework for rearrangement in atomic arrays

The task of atom rearrangement has emerged in the last decade as a fundamental building block in the development of neutral atom-based quantum processors. As such processors grow to thousands of atoms, it becomes increasingly important to design algorithms robust to experimental sources of error. While recent progress has been made towards developing algorithms with favorable time scaling, such work has been limited to noiseless settings. Moreover, there is a lack of open-source code for reproducing and benchmarking existing algorithms. To address these deficiencies, we develop an open-source simulation framework, atommovr, and leverage it to study three distinct settings: 1) time-optimal, noiseless rearrangement, 2) noisy rearrangement under realistic error models, and 3) noiseless dual-species rearrangement. We extract lower bounds for time-optimal rearrangement, study advantageous strategies across different error regimes, and develop a novel dual-species algorithm, InsideOut, capable of avoiding 'blocked' configurations with a near-unity success rate. We hope that atommovr can serve as a common tool for the community to study rearrangement, lower the barrier to entry for new experimental groups, and stimulate progress in developing algorithms tailored to minimize atom loss in experiment.