Mixed-Timescale Online PHY Caching for Dual-Mode MIMO Cooperative Networks
Mixed-Timescale Online PHY Caching for Dual-Mode MIMO Cooperative Networks
复制标题
DOI：
10.1109/twc.2019.2907586
复制
发表时间：
2018-10
影响因子：
10.4
通讯作者：
An Liu;V. Lau;Wenchao Ding;E. Yeh
中科院分区：
文献类型：
--
作者：
An Liu;V. Lau;Wenchao Ding;E. Yeh
近年来，物理层（PHY）缓存被提出用于利用基站（BSs）缓存引起的动态侧信息来支持协调多点（CoMP）并获得高自由度（DoF）增益。由于缓存存储容量有限，PHY缓存的性能在很大程度上取决于缓存内容放置算法。在现有算法中，缓存内容的放置以离线方式适应长期流行度分布。我们提出了一个在线PHY缓存框架，该框架使缓存内容的放置适应微观空间和临时流行度的变化，以充分利用PHY缓存的优势。具体来说，在线缓存内容放置和内容交付的联合优化被制定为一个混合时间尺度漂移最小化问题，以增加CoMP机会并降低缓存内容放置成本。我们提出了一种低复杂度的算法来获得吞吐量最优解。此外，我们还提供了稳定区域最大和自由度的封闭形式表征，并研究了系统关键参数对稳定区域的影响。仿真结果表明，所提出的在线物理缓存框架比现有的解决方案获得了较大的增益。
Recently, physical layer (PHY) caching has been proposed to exploit the dynamic side information induced by caches at base stations (BSs) to support coordinated multi-point (CoMP) and achieve high degrees of freedom (DoF) gains. Due to the limited cache storage capacity, the performance of PHY caching depends heavily on the cache content placement algorithm. In the existing algorithms, the cache content placement is adaptive to the long-term popularity distribution in an offline manner. We propose an online PHY caching framework, which adapts the cache content placement to microscopic spatial and temporary popularity variations to fully exploit the benefits of PHY caching. Specifically, the joint optimization of online cache content placement and content delivery is formulated as a mixed-timescale drift minimization problem to increase the CoMP opportunity and reduce the cache content placement cost. We propose a low-complexity algorithm to obtain a throughput-optimal solution. Moreover, we provide a closed-form characterization of the maximum sum DoF in the stability region and study the impact of key system parameters on the stability region. The simulations results show that the proposed online PHY caching framework achieves large gain over existing solutions.