Quantum simulation costs for Suzuki-Trotter decomposition of quantum many-body lattice models

TL;DR

This paper develops a formalism to estimate the number of Trotter steps needed for accurate quantum simulation of fermionic lattice models, revealing that the t-J model may be more efficiently simulated than the Hubbard model under realistic conditions.

Contribution

The paper introduces a new formalism for bounding Trotter step counts in quantum simulations of many-body lattice models, aiding in cost assessment for near-term quantum computers.

Findings

The formalism provides bounds based on first-order commutator scaling.

Careful parameter analysis favors the t-J model over the Hubbard model.

Results suggest potential for reducing quantum simulation costs.

Abstract

Quantum computers offer the potential to efficiently simulate the dynamics of quantum systems, a task whose difficulty scales exponentially with system size on classical devices. To assess the potential for near-term quantum computers to simulate many-body systems we develop a formalism to straightforwardly compute bounds on the number of Trotter steps needed to accurately simulate the time evolution of fermionic lattice models based on the first-order commutator scaling. We apply this formalism to two closely related many-body models prominent in condensed matter physics, the Hubbard and t-J models. We find that, while a naive comparison of the Trotter depth first seems to favor the Hubbard model, careful consideration of the model parameters and the allowable error for accurate simulation leads to a substantial advantage in favor of the t-J model. These results and formalism set the stage for significant improvements in quantum simulation costs.