Variational Quantum Optimization of Nonlocality in Noisy Quantum Networks

Abstract: The inherent noise and complexity of quantum communication networks leads to challenges in designing quantum network protocols using classical methods. To address this issue, we develop a variational quantum optimization framework that simulates quantum networks on quantum hardware and optimizes the network using differential programming techniques. We use our hybrid framework to optimize nonlocality in noisy quantum networks. On the noisy IBM quantum computers, we demonstrate our framework's ability to maximize quantum nonlocality. On a classical simulator with a static noise model, we investigate the noise robustness of quantum nonlocality with respect to unital and nonunital channels. In both cases, we find that our optimization methods can reproduce known results, while uncovering interesting phenomena. When unital noise is present we find numerical evidence suggesting that maximally entangled state preparations yield maximal nonlocality. When nonunital noise is present we find that nonmaximally entangled states can yield maximal nonlocality. Thus, we show that variational quantum optimization is a practical design tool for quantum networks in the near-term. In the long-term, our variational quantum optimization techniques show promise of scaling beyond classical approaches and can be deployed on quantum network hardware to optimize quantum communication protocols against their inherent noise.