This paper studies caching in $(K\!+\!L\!-\!1)\!\times \!K$ partially connected wireless linear networks, where each of the $K$ receivers locally communicates with $L$ out of the $K\!+\!L\!-\!1$ transmitters, and caches are at all nodes. The goal is to design caching and delivery schemes to reduce the transmission latency, by using normalized delivery time (NDT) as the performance metric. For small transmitter cache size (any $L$ transmitters can collectively store the database just once), we propose a cyclic caching strategy so that each of every $L$ consecutive transmitters caches a distinct part of each file; the delivery strategy exploits coded multicasting and interference alignment by introducing virtual receivers. The obtained NDT is within a multiplicative gap of 2 to the optimum in the entire cache size region, and optimal in certain region. For large transmitter cache size (any $L$ transmitters can collectively store the database for multiple copies), we propose a modified caching strategy so that every bit is repeatedly cached at consecutive transmitters; the delivery strategy exploits self-interference cancellation and interference neutralization. By combining these schemes, the NDT is optimal in a larger region. We also extend our results to linear networks with heterogeneous receiver connectivity and partially connected circular networks.

中文翻译：

---

部分连接线性网络中的缓存辅助干扰管理

本文研究缓存 $(K\!+\!L\!-\!1)\!\times \!K$ 部分连接的无线线性网络，其中每个 $K$ 接收器与本地通信 $L$ 出于 $K\!+\!L\!-\!1$ 发送器和缓存位于所有节点。目标是通过使用归一化交付时间 (NDT) 作为性能指标来设计缓存和交付方案以减少传输延迟。对于较小的发射器缓存大小（任何 $L$ 发送器可以共同存储数据库一次），我们提出了一种循环缓存策略，以便每个 $L$ 连续发送器缓存每个文件的不同部分；交付策略通过引入虚拟接收器来利用编码多播和干扰对齐。得到的无损检测在整个缓存大小区域内与最优值在 2 倍的乘法间隙内，并且在某个区域是最优的。对于大型发射器缓存大小（任何 $L$ 发送器可以共同存储多个副本的数据库），我们提出了一种修改的缓存策略，以便在连续的发送器处重复缓存每个位；交付策略利用自干扰消除和干扰抵消。通过组合这些方案，无损检测在更大的区域内是最佳的。我们还将我们的结果扩展到具有异构接收器连接性和部分连接的圆形网络的线性网络。