Variational quantum eigensolver techniques for simulating carbon monoxide oxidation

Mariia D. Sapova

doi:10.1038/s42005-022-00982-4

Return

Variational quantum eigensolver techniques for simulating carbon monoxide oxidation

2022-08-06

Mariia D. Sapova

Aleksey K. Fedorov *

DOI：10.1038/s42005-022-00982-4

Original

Original request for help

Save

Abstract

En 中文

The variational quantum eigensolver is a quantum-classical algorithm used to solve optimisation problems in machine learning but demonstrates limitations when applied to simulations of large molecules. Here, the authors explore the use of adaptive variational algorithms and demonstrate how they can be used to improve performance when simulating molecules participating in carbon monoxide processes. Variational Quantum Eigensolver (VQE) methods aim to maximize the resources of existing noisy devices. However, they encounter difficulties in simulating molecules of industrially-relevant sizes, such as constructing the efficient ansatz. Adaptive variational algorithms (ADAPT-VQE) can solve this problem but with a significant increase in the number of measurements. Here, we reduce the measurement overhead of ADAPT-VQE via adding operator batches to the ansatz while keeping it compact. We reformulate the previously proposed qubit pool completeness criteria for the tapered qubit space and propose an automated pool construction procedure. Our numerical results indicate that reducing the qubit pool size from polynomial to linear conversely increases the number of measurements. We simulate a set of molecules, participating in the carbon monoxide oxidation processes using the statevector simulator and compare the results with VQE-UCCSD and classical methods. Our results pave the way towards usage of variational approaches for solving practically relevant chemical problems.

Keywords:

COUPLED-CLUSTER