IEEE Transactions on Networking
Efficient Quantum Link Selection and Fidelity Estimation Under Success Rate Constraints
Maoli Liu, Zhuohua Li, Xuchuang Wang, John C. S. Lui
Abstract
Quantum networks enable secure communication across distributed quantum systems. To reliably transmit quantum information, entanglement links need to exhibit both highfidelityand sufficientsuccess rates. However, these crucial metrics are initially unknown, and estimating them uniformly across all links in large-scale quantum networks is highly resource-intensive. In this work, we address the problem of efficiently identifying the link with the highest fidelity among those satisfying a minimum success rate constraint. We formulate it as a constrained best arm identification problem and introduce two algorithms, LINKSELFIE-E and LINKSELFIE-C, built upon thenetwork benchmarkingsubroutine. To enable joint fidelity and success rate estimation, we extend network benchmarking by incorporating a minimum-variance unbiased estimator for success rates. LINKSELFIE-E is an elimination-based algorithm that iteratively removes inferior links in successive phases, while LINKSELFIE-C is a confidence-based algorithm with an interactive update mechanism, dynamically refining estimates and prioritizing benchmarking resources toward promising links. We derive confidence intervals for fidelity and success rate estimates and prove that both LINKSELFIE-E and LINKSELFIE-C identify the optimal link with high probability. We also establish upper bounds on their cost complexities, demonstrating their efficiency. Extensive simulations under various noise models show that our algorithms outperform baseline methods in both identifying the optimal link and reducing quantum resource consumption.