First-Order Trotter Error from a Second-Order Perspective

Simulating quantum dynamics beyond the reach of classical computers is one of the main envisioned applications of quantum computers. The most promising quantum algorithms to this end in the near term are the simplest, which use the Trotter formula and its higher-order variants to approximate the dynamics of interest. The approximation error of these algorithms is often poorly understood, even in the most basic and topical cases where the target Hamiltonian decomposes into two realizable terms: H 1/4 H1 thorn H2. Recent studies have reported anomalously low approximation error with unexpected scaling in such cases, which they attribute to quantum interference between the errors from different steps of the algorithm. Here, we provide a simpler picture of these effects by relating the Trotter formula to its second-order variant for such H 1/4 H1 thorn H2 cases. Our method generalizes state-of-the-art error bounds without the technical caveats of prior studies, and elucidates how each part of the total error arises from the underlying quantum circuit. We compare our bound to the true error numerically, and find a close match over many orders of magnitude in the simulation parameters. Our findings further reduce the required circuit depth for the least experimentally demanding quantum simulation algorithms, and illustrate a useful method for bounding simulation error more broadly.

Automated summary

Simulating quantum dynamics beyond classical computers' capabilities is a major application of quantum computers. The Trotter formula and its higher-order variants are the most promising quantum algorithms for this purpose in the near future. However, the approximation error of these algorithms is often poorly understood, especially in cases where the target Hamiltonian can be decomposed into two realizable terms. This study provides a simpler understanding of these effects by relating the Trotter formula to its second-order variant for such cases. The findings reduce the required circuit depth for quantum simulation algorithms and illustrate a useful method for bounding simulation error more broadly.