This paper tests a quantum algorithmic approach to prepare the lowest-energy state of a two-dimensional Z2 lattice gauge theory. The authors apply a deterministic quantum imaginary-time evolution (QITE) method and adapt it so the algorithm respects the model's local gauge constraints (Gauss's law). They then check how well the method works by running classical, noiseless simulations and comparing to established numerical methods.

At the algorithmic level they generalize a construction of Pauli operators that commute with Gauss's law to arbitrary supports. These Pauli operators form the building blocks of the QITE unitary approximation. Deterministic QITE approximates non-unitary imaginary-time evolution by a sequence of unitary evolutions built from Pauli strings. The coefficients of those Pauli strings are found by solving a linear equation. The authors use a second-order Suzuki-Trotter decomposition to split the total evolution into many small steps and then approximate each small imaginary-time piece by a real-time unitary.

For testing, the authors ran classical simulations with tensor-network techniques and compared the results to density matrix renormalization group (DMRG) and to a Suzuki-Trotterized imaginary-time evolution without the unitary approximation. They focused on a ladder-like two-dimensional geometry (two links in the vertical direction) and varied the system size in the horizontal direction and the coupling strength. In the regime they studied, deterministic QITE reached a relative energy error below 0.1 percent for systems up to twelve plaquettes. They also investigated how the error depends on the number of time steps.

Why this matters: preparing ground states is a key step for using lattice gauge models to study quantum phases and dynamics. Making the QITE procedure gauge-invariant by restricting to Pauli operators that commute with Gauss's law both preserves the physical symmetry and cuts the number of measurements and quantum gates needed. The paper shows that, for modest two-dimensional systems and the chosen geometry, deterministic QITE can be highly accurate when implemented idealistically.