This paper considers a hierarchical content delivery network in a randomly distributed interference environment, where the caching capability is distributed among two layers, the caching nodes and part of users. We propose a hierarchical caching strategy that is jointly optimized with the delivery performance, to achieve a fine balance between the cooperative gain and the caching diversity gain. In particular, to guarantee the effectiveness of caching strategy, the impact of communication interference at different layers on caching strategy is discussed during the file delivery phase. The optimization problem to maximize the offloading probability is formulated as a combinatorial optimization problem. By breaking the problem into two sub-multiple choice knapsack problems, the extreme case in the absence of user caching ability is first analyzed, and an optimization method based on the worst-case among users is proposed. Detailed analytical analysis and simulation experiments are carried out to evaluate the proposed caching strategy. The experimental results prove the superiority of the proposed cache strategy by comparing it with other caching strategies.

本文考虑随机分布干扰环境下的分层内容交付网络，其中缓存能力分布在缓存节点和部分用户两层之间。我们提出了一种与交付性能联合优化的分层缓存策略，以实现协作增益和缓存多样性增益之间的良好平衡。特别是为了保证缓存策略的有效性，在文件传递阶段讨论了不同层级的通信干扰对缓存策略的影响。最大化卸载概率的优化问题被表述为组合优化问题。通过将问题分解为两个子多项选择背包问题，首先分析没有用户缓存能力的极端情况，并提出了一种基于用户间最坏情况的优化方法。进行详细的分析分析和模拟实验以评估所提出的缓存策略。通过与其他缓存策略的比较，实验结果证明了所提出的缓存策略的优越性。