A Parallel Genetic Algorithm for Qubit Mapping on Noisy Intermediate-Scale Quantum Machines
Résumé
Quantum computers are getting increasingly large and available thanks to some major technological advancements, but they remain in the realm of NISQ (Noisy Intermediate Scale Quantum) devices. On such devices, due to the limited connectivity of the physical qubits, most quantum circuit-based programs cannot be executed without transpilation. This latter includes an important step, referred to as qubit mapping, which consists in converting the quantum circuit into another one which best matches the graph of physical qubits taking into account its limited connectivity constraint. In this paper, we propose a Parallel Genetic Algorithm to Qubit Mapping (PGA-QM). The challenge is to minimize the depth of the transformed circuit and the execution time and error rate consequently. PGA-QM has been experimented using various medium-to-large scale circuit benchmarks. It is compared against the SABRE heuristic currently implemented in Qiskit, the IBM's library for quantum computing. The reported results show that PGA-QM can provide better solutions and with better consistency than its counterpart while parallelism greatly reduces its execution time during the transpilation.
Origine | Fichiers produits par l'(les) auteur(s) |
---|---|
licence |
Copyright (Tous droits réservés)
|