2020/08/13 by Jesper Hasseriis Mohr Jensen, Frederik Møller, Jensen, Jesper Hasseriis Mohr +5 · 1 citation
Physics and Astronomy · #Atomic and Subatomic Physics Research #Cold Atom Physics and Bose-Einstein Condensates #FOS: Physical sciences #Physics of Superconductivity and Magnetism #Quantum Physics (quant-ph)
paper · pdf · doi:10.48550/arxiv.2008.06076
openalex publication_date 2020/08/13 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
We demonstrate the efficiency of a recent exact-gradient optimal control\nmethodology by applying it to a challenging many-body problem, crossing the\nsuperfluid to Mott-insulator phase transition in the Bose-Hubbard model. The\nsystem size necessitates a matrix product state representation and this\nseamlessly integrates with the requirements of the algorithm. We observe\nfidelities in the range 0.99-0.9999 with associated minimal process duration\nestimates displaying an exponential fidelity-duration trade-off across several\norders of magnitude. The corresponding optimal solutions are characterized in\nterms of a predominantly linear sweep across the critical point followed by\nbang-bang-like structure. This is quite different from the smooth and monotonic\nsolutions identified by earlier gradient-free optimizations which are hampered\nin locating the higher complexity protocols in the regime of high-fidelities at\nlow process durations. Overall, the comparison suggests significant\nmethodological improvements also for many-body systems in the ideal open-loop\nsetting. Acknowledging that idealized open-loop control may deteriorate in\nactual experiments, we discuss the merits of using such an approach in\ncombination with closed-loop control -- in particular, high-fidelity physical\ninsights extracted with the former can be used to formulate practical,\nlow-dimensional search spaces for the latter.\n