Mitigating algorithmic errors in quantum optimization through energy extrapolation

Abstract

Quantum optimization algorithms offer a promising route to finding the ground states of target Hamiltonians on near-term quantum devices. Nonetheless, it remains necessary to limit the evolution time and circuit depth as much as possible, since otherwise decoherence will degrade the computation. Even when this is done, there always exists a non-negligible error in estimates of the ground state energy. Here we present a scalable extrapolation approach to mitigating this algorithmic error, which significantly improves estimates obtained using three well-studied quantum optimization algorithms: quantum annealing (QA), the variational quantum eigensolver, and the quantum imaginary time evolution at fixed evolution time or circuit depth. The approach is based on extrapolating the annealing time to infinity or the variance of estimates to zero. The method is reasonably robust against noise. For Hamiltonians which only involve few-body interactions, the additional computational overhead is an increase in the number of measurements by a constant factor. Analytic derivations are provided for the quadratic convergence of estimates of energy as a function of time in QA, and the linear convergence of estimates as a function of variance in all three algorithms. We have verified the validity of these approaches through both numerical simulation and experiments on IBM quantum machines. This work suggests a promising new way to enhance near-term quantum computing through classical post-processing.