Perturbative Variational Quantum Eigensolver based on Reduced Density Matrix Method

Abstract: Current noisy intermediate-scale quantum (NISQ) devices lack the quantum resources required for practical applications. To address this, we propose the perturbative variational quantum eigensolver (PT-VQE). In PT-VQE, VQE is used to capture key correlations within a carefully selected active space, while perturbation theory efficiently incorporates interactions in the remaining space, without requiring additional qubits or circuit depth. When the VQE-optimized state closely approximates the true ground state in the active space, excitations cannot act solely in the active space, since their contributions to perturbative correction are negligible. This reduces the highest-order required RDM from 4-RDM to 3-RDM, significantly reducing computational costs. We validate PT-VQE by calculating the ground-state potential energy surfaces (PESs) of $\rm{HF}$ and $\rm{N}_2$, as well as the ground-state energy of ferrocene ($\rm{Fe(C_5H_5)_2}$). Additionally, PT-VQE is performed on a quantum computer to compute the PES of ${\rm F}_2$. The consistent results obtained from both PT-VQE with the highest 3-RDM and 4-RDM confirm the reliability of the constraint. PT-VQE significantly outperforms standard VQE, achieving chemical accuracy. This method offers a resource-efficient and practical approach for accurate quantum simulations of larger systems.