Given the diverse array of physical systems available for quantum computing and the absence of a well-defined quantum internet protocol stack, the design and optimisation of quantum networking protocols remain largely unexplored. To address this, we introduce an approach that facilitates the establishment of paths capable of delivering end-to-end fidelity above a specified threshold, without requiring detailed knowledge of the quantum network properties, which we call the 'grey box approach'. In this study, we define algorithms that are specific instances of this approach and evaluate them in comparison to Dijkstra shortest path algorithm and a fully knowledge-aware algorithm through simulations. Our results demonstrate that one of the grey box algorithms consistently outperforms the other methods in delivering paths above the fidelity threshold, across various network topologies and the number of source-destination pairs involved, while maintaining significant levels of fairness among the users and being robust to inaccurate estimations of the expected end-to-end fidelity.

翻译：鉴于可用于量子计算的物理系统种类繁多，且缺乏明确定义的量子互联网协议栈，量子网络协议的设计与优化在很大程度上仍未被探索。为解决这一问题，我们提出了一种方法，该方法能够促进建立能够提供高于指定阈值的端到端保真度路径，而无需详细了解量子网络特性，我们称之为"灰盒方法"。在本研究中，我们定义了作为该方法具体实例的算法，并通过仿真将其与Dijkstra最短路径算法和完全知识感知算法进行比较评估。我们的结果表明，在各种网络拓扑结构和所涉及的源-目的对数量下，其中一种灰盒算法在提供高于保真度阈值的路径方面始终优于其他方法，同时保持用户间显著水平的公平性，并且对预期端到端保真度的不准确估计具有鲁棒性。