Comparing Classical and Quantum Ground State Preparation Heuristics

TL;DR

This paper compares classical and quantum heuristic methods for preparing ground states in quantum systems, showing quantum approaches can improve efficiency and reduce runtime in energy estimation tasks.

Contribution

It demonstrates that quantum heuristic ground state preparation can outperform classical methods in certain cases, especially for intermediate-sized strongly-correlated systems.

Findings

Quantum heuristics improve overlap in ground state preparation.

Quantum methods accelerate ground state energy estimation.

Potential for reduced quantum resource requirements.

Abstract

One promising field of quantum computation is the simulation of quantum systems, and specifically, the task of ground state energy estimation (GSEE). Ground state preparation (GSP) is a crucial component in GSEE algorithms, and classical methods like Hartree-Fock state preparation are commonly used. However, the efficiency of such classical methods diminishes exponentially with increasing system size in certain cases. In this study, we investigated whether in those cases quantum heuristic GSP methods could improve the overlap values compared to Hartree-Fock. Moreover, we carefully studied the performance gain for GSEE algorithms by exploring the trade-off between the overlap improvement and the associated resource cost in terms of T-gates of the GSP algorithm. Our findings indicate that quantum heuristic GSP can accelerate GSEE tasks, already for computationally affordable strongly-correlated systems of intermediate size. These results suggest that quantum heuristic GSP has the potential to significantly reduce the runtime requirements of GSEE algorithms, thereby enhancing their suitability for implementation on quantum hardware.