In this article, delay and energy efficiency (EE) are investigated in fog radio access networks (F-RANs) with hybrid caching. With multiple caching and transmission strategies, hybrid caching offers great flexibility for file placement and file fetching. By using tools from stochastic geometry, we firstly derive tractable expressions of delay for coded cached, non-partitioned cached and uncached files. Then, we derive tractable expressions of EE by jointly considering power consumed in circuits, transmissions and fronthaul links. To balance delay and EE, the corresponding multi-objective optimization problem is formulated to obtain the optimal hybrid caching strategy. Furthermore, considering the NP-hard complexity of the problem, we first theoretically analyze the optimal structure of the caching result. Then, we convert the original problem into a classification problem. We further propose a gradual-replacement greedy algorithm to obtain a near optimal hybrid caching strategy, which ensures high accuracy with low complexity. Numerical results show a significant performance gain of the proposed near optimal hybrid caching strategy over baselines and flexibility in delay-sensitive and EE-sensitive scenarios.

中文翻译：

---

具有混合缓存的雾无线接入网络的分析和优化：延迟和能源效率

在本文中，研究了具有混合缓存的雾无线电接入网络 (F-RAN) 中的延迟和能源效率 (EE)。通过多种缓存和传输策略，混合缓存为文件放置和文件获取提供了极大的灵活性。通过使用来自随机几何的工具，我们首先为编码的缓存、非分区缓存和未缓存文件推导出易于处理的延迟表达式。然后，我们通过联合考虑电路、传输和前传链路中消耗的功率来推导出易于处理的 EE 表达式。为了平衡延迟和EE，制定相应的多目标优化问题以获得最佳混合缓存策略。此外，考虑到问题的NP-hard复杂性，我们首先从理论上分析了缓存结果的最优结构。然后，我们将原始问题转化为分类问题。我们进一步提出了一种渐进替换贪婪算法，以获得接近最优的混合缓存策略，确保高精度和低复杂度。数值结果表明，在延迟敏感和 EE 敏感场景中，所提出的近乎最优混合缓存策略在基线和灵活性方面具有显着的性能提升。