Orbital-rotated Fermi-Hubbard model as a benchmarking problem for quantum chemistry with the exact solution

Evaluating the relative performance of different quantum algorithms for quantum computers is of great significance in the research of quantum algorithms. In this study, we consider the problem of quantum chemistry, which is considered one of the important applications of quantum algorithms. While evaluating these algorithms in systems with a large number of qubits is essential to see the scalability of the algorithms, the solvable models usually used for such evaluations typically have a small number of terms compared to the molecular Hamiltonians used in quantum chemistry. The large number of terms in molecular Hamiltonians is a major bottleneck when applying quantum algorithms to quantum chemistry. Various methods are being considered to address this problem, highlighting its importance in developing quantum algorithms for quantum chemistry. Based on these points, a solvable model with a number of terms comparable to the molecular Hamiltonian is essential to evaluate the performance of such algorithms. In this paper, we propose a set of exactly solvable Hamiltonians that has a comparable order of terms with molecular Hamiltonians by applying a spin-involving orbital rotation to the one-dimensional Fermi-Hubbard Hamiltonian. We verify its similarity to the molecular Hamiltonian from some prospectives and investigate whether the difficulty of calculating the ground-state energy changes before and after orbital rotation by applying the density matrix renormalization group up to 24 sites corresponding to 48 qubits. This proposal would enable proper evaluation of the performance of quantum algorithms for quantum chemistry, serving as a guiding framework for algorithm development.
Comment: 10 pages