Work

Variaatiolaskentaa hyödyntävien kvnattialgoritmien vertailututkimus pienille molekyyleille

Year: 2024
Type: article
Abstract: Finding the ground state energy of a molecule is a fundamental task in quantum chemistry. This process typically involves solving for the lowest eigenstate of a Hamiltonian matrix, whose dimensions scale as a function of the number of active spin-orbitals. Solving this problem is exponentially demanding on classical computers, restricting exact solutions to small- and medium-sized molecules. However, quantum computers can theoretically perform this task more efficiently using a quantum algorithm known as Quantum Phase Estimation (QPE). Unfortunately, current quantum computers are noisy and lack sufficient physical qubits to implement QPE effectively. This era of quantum computing, marked by noise and limited qubit counts, is referred to as the Noisy Intermediate-Scale Quantum (NISQ) era. To address these hardware limitations, a hybrid algorithm called the Variational Quantum Eigen- solver (VQE) was developed in 2014. VQE significantly reduces the hardware requirements for solving the ground state energy problem compared to QPE. For VQE to achieve a quantum advantage on NISQ-era quantum computers, careful selection of implementation methods is essential. Poor implementation can negate any quantum advantage the algorithm might offer. Key considerations in VQE include choosing the initial state ansatz, ansatz parametrization, fermion-to-Pauli mapping, and measurement scheme. The selection and optimization of these components are complex and remain active areas of research. We compared these methods on molecules with between four and 24 spin-orbitals. Our findings indicate that the Bravyi-Kitaev fermion-to-Pauli encoding is the most effective for systems with 16 or more spin-orbitals in terms of accuracy. For observables, the Jordan-Wigner encoding and Qubit-Wise Commuting (QWC) measurement scheme resulted in the lowest count—31% fewer than Bravyi-Kitaev at 24 spin-orbitals. Overall, we found that a combination of the QWC measurement scheme and hardware-efficient ansatz parametrization performed best, providing a scalable speed-up in orders of magnitude in a noisy emulation of a quantum computer compared to similar VQE runs using the Pauli measurement scheme and Jordan-Wigner encoding. This speed-up was observed across all simulated molecules and was particularly notable for water, where computational time was reduced from 28 hours to just 10 minutes.
Language: English
Open Access status: green
OpenAlex ID: W7197638572
FWCI:
Cites:
Cited by:
Related to: 10