Evaluating performance of hybrid quantum optimization algorithms for MAXCUT Clustering using IBM runtime environment
Abstract
Quantum algorithms can be used to perform unsupervised machine learning tasks like data clustering by mapping the distance between data points to a graph optimization problem (i.e. MAXCUT) and finding optimal solution through energy minimization using hybrid quantum classical methods. Taking advantage of the IBM runtime environment, we benchmark the performance of the "Warm-Start" (ws) variant of Quantum Approximate Optimization Algorithm (QAOA) versus the standard implementation of QAOA and the variational quantum eigensolver (VQE) for unstructured clustering problems using real world dataset with respect to accuracy and execution time. Our numerical results show a strong speedup in execution time for different optimization algorithms using the IBM Qiskit Runtime architecture and increased speedup in classification accuracy in ws-QAOA algorithm
Keywords
Cite
@article{arxiv.2112.03199,
title = {Evaluating performance of hybrid quantum optimization algorithms for MAXCUT Clustering using IBM runtime environment},
author = {Daniel Beaulieu and Anh Pham},
journal= {arXiv preprint arXiv:2112.03199},
year = {2022}
}
Comments
arXiv admin note: substantial text overlap with arXiv:2108.13464