arXiv:2610.08667
Fault-tolerant resource estimation for ground-state preparation via Lindblad simulation
Marius Bothe, Patrick Schoepf, Nick S. Blunt
Abstract
Recent advances in algorithms for simulating Lindblad dynamics have clarified their theoretical potential for state preparation, but their practicality in the early fault-tolerant regime and beyond remains uncertain. In this work, we address this by investigating the cost of preparing ground states of the fermionic Hubbard model, following the single-ancilla approach of [Phys. Rev. Research 6, 033147 (2024)]. We derive rigorous error bounds including constant prefactors, and compare to empirical error behavior and practical convergence parameters obtained from circuit-level simulations. We find that empirically chosen parameters can reduce the required resources by orders of magnitude, resulting in an estimated $7.7 \times 10^8$ T gates to perform one unit of time evolution targeting the low-energy subspace of a 36-site fermionic Hubbard model. We identify the accurate filtering of energy transitions as the main source of this cost.
AI Summary
The paper asks what fault-tolerant resources are required for ground-state preparation using single-ancilla, dilation-based Lindblad simulation. It studies filtered, particle-number- and spin-projection-preserving hopping operators for the fermionic Hubbard model, compiling the dynamics into Trotterized Hamiltonian evolutions, ancilla interactions, measurement/reset, and Clifford+$T$ rotation synthesis. The analysis supplies explicit-prefactor trace-norm bounds for Lindblad splitting, dilation, filter quadrature, and product-formula errors. For the bump filter, the quadrature tail decays as $\exp[-\sqrt{a\pi(M_S-1/2)}/6]$, while resolving a gap $\Delta$ requires a filtering window scaling inversely with $\Delta$. Circuit-level simulations of half-filled, open 1D Hubbard chains at $U/t_{\mathrm{hop}}=4$, up to eight sites, provide empirical convergence parameters. These reduce estimated costs by several orders of magnitude relative to rigorous error bounds. For 36 sites, one unit of Lindblad evolution costs approximately $7.7\times10^8$ $T$ gates using general empirical parameters, versus $3.3\times10^{15}$ with rigorous parameters; a gap-informed empirical regime gives $1.2\times10^8$. The implementations require 74 logical qubits. Hamiltonian evolution inside the energy-filter integral dominates the cost. These are per-unit-time estimates, not complete preparation costs: large-system mixing times remain unknown, and the stated step-count scaling is quadratic in total evolution time. Empirical extrapolations rely on small 1D systems; similar 2D per-time costs are argued rather than simulated. The target accuracy is an extensive energy error, $0.005n_{\mathrm{sites}}$, rather than guaranteed ground-state fidelity. Noise simulations also show that symmetry-sector leakage is not repaired by the dissipative dynamics.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.