Quantum simulation of the 1D Fermi-Hubbard model as a $\mathrm{Z}_2$ lattice-gauge theory

Uliana E. Khodaeva,Dmitry L. Kovrizhin,Johannes Knolle
2023-11-08
Abstract:The Fermi-Hubbard model is one of the central paradigms in the physics of strongly-correlated quantum many-body systems. Here we propose a quantum circuit algorithm based on the $\mathrm{Z}_2$ lattice gauge theory (LGT) representation of the one-dimensional Fermi-Hubbard model, which is suitable for implementation on current NISQ quantum computers. Within the LGT description there is an extensive number of local conserved quantities commuting with the Hamiltonian. We show how these conservation laws can be used to implement an efficient error-mitigation scheme. The latter is based on a post-selection of states for noisy quantum simulators. While the LGT description requires a deeper quantum-circuit compared to a Jordan-Wigner (JW) based approach, remarkably, we find that our error-correction protocol leads to results being on-par with a standard JW implementation on noisy quantum simulators.
Quantum Physics,Strongly Correlated Electrons
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to efficiently simulate the one - dimensional Fermi - Hubbard model (1D Fermi - Hubbard model) on current noisy intermediate - scale quantum (NISQ) computers. Specifically, the authors propose a quantum circuit algorithm based on \(Z_2\) lattice gauge theory (LGT) to achieve an effective quantum simulation of the 1D Fermi - Hubbard model. This method focuses particularly on how to use a large number of local conserved quantities in the LGT description to implement an effective error - mitigation scheme. ### Main problems 1. **Effectively simulate the 1D Fermi - Hubbard model**: - The Fermi - Hubbard model is a central paradigm for strongly - correlated quantum many - body systems, but its exact solutions are limited to certain specific points in the parameter space. Achieving an efficient quantum simulation of this model on NISQ computers is a challenge. 2. **Error mitigation**: - NISQ computers have inevitable noise and errors, which will affect the accuracy of the simulation results. The paper proposes a post - selection error - mitigation scheme based on local conserved quantities to improve the reliability of the simulation. ### Solutions 1. **LGT representation**: - Map the 1D Fermi - Hubbard model to a \(Z_2\) lattice gauge theory by introducing auxiliary spin - 1/2 degrees of freedom to describe the coupling between fermions on the lattice and spin - 1/2 degrees of freedom on the links. - This mapping allows the use of a large number of local conserved quantities that remain invariant during the time - evolution process. 2. **Quantum circuit implementation**: - Design a quantum circuit based on \(Z_2\) LGT, using CNOT gates and other single - qubit gates to implement time - evolution. - Although the LGT method requires more CNOT gates (six CNOT gates are required for each hopping term, while only two are required for the direct method), through the post - selection error - mitigation scheme, the performance of the LGT method on a noisy quantum simulator can be comparable to or even better than that of the standard Jordan - Wigner transformation method. 3. **Error - mitigation strategies**: - **Post - selection**: Use local conserved quantities for post - selection and ignore those measurement results that violate the conserved quantities. - **Normalization**: Measure the decay of local operators by reducing the Trotter step size and correct the noise influence by normalization. ### Experimental results - **Noiseless case**: On an ideal quantum simulator, the Trotterized time - evolution results of the LGT method are very close to the exact diagonalization (ED) results. - **Noisy case**: On Qiskit's QasmSimulator, through the post - selection and normalization strategies, the performance of the LGT method at different noise levels is significantly better than that without error - mitigation. Even under strong noise (\(\eta = 0.01\)), the results of the LGT method are still relatively good. ### Conclusion The paper demonstrates the potential of the quantum circuit algorithm based on \(Z_2\) lattice gauge theory for simulating the 1D Fermi - Hubbard model on NISQ computers, especially the advantages in error - mitigation. Although the LGT method requires more quantum resources, through effective error - mitigation strategies, its performance can be comparable to or even better than that of the traditional Jordan - Wigner transformation method. This research provides new ideas for further exploring the quantum simulation of strongly - correlated systems.