Quantum circuit-width reduction through parameterisation and specialisation

Moawad, Y., Vanderbauwhede, W. and Steijl, R. (2023) Quantum circuit-width reduction through parameterisation and specialisation. Algorithms, 16(5), 241. (doi: 10.3390/a16050241)

Abstract

As quantum computing technology continues to develop, the need for research into novel quantum algorithms is growing. However, such algorithms cannot yet be reliably tested on actual quantum hardware, which is still limited in several ways, including qubit coherence times, connectivity, and available qubits. To facilitate the development of novel algorithms despite this, simulators on classical computing systems are used to verify the correctness of an algorithm, and study its behaviour under different error models. In general, this involves operating on a memory space that grows exponentially with the number of qubits. In this work, we introduce quantum circuit transformations that allow for the construction of parameterised circuits for quantum algorithms. The parameterised circuits are in an ideal form to be processed by quantum compilation tools, such that the circuit can be partially evaluated prior to simulation, and a smaller specialised circuit can be constructed by eliminating fixed input qubits. We show significant reduction in the number of qubits for various quantum arithmetic circuits. Divide-by-n-bits quantum integer dividers are used as an example demonstration. It is shown that the complexity reduces from 4n+2 to 3n+2 qubits in the specialised versions. For quantum algorithms involving divide-by-8 arithmetic operations, a reduction by 28=256 in required memory is achieved for classical simulation, reducing the memory required from 137 GB to 0.53 GB.

• Item Type: Articles
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Steijl, Dr Rene and Vanderbauwhede, Professor Wim and Moawad, Youssef
• Authors: Moawad, Y., Vanderbauwhede, W., and Steijl, R.
• College/School: College of Science and Engineering > School of Computing Science; College of Science and Engineering > School of Engineering > Autonomous Systems and Connectivity
• Journal Name: Algorithms
• Publisher: MDPI
• ISSN: 1999-4893
• ISSN (Online): 1999-4893
• Published Online: 05 May 2023
• Copyright Holders: Copyright © 2023 The Authors
• First Published: First published in Algorithms 16(5): 241
• Publisher Policy: Reproduced under a Creative Commons License
• Data DOI: 10.5281/zenodo.7645972