In this work, we study the generalized degrees-of-freedom (GDoF) of downlink and uplink cellular networks, modeled as Gaussian interfering broadcast channels (IBC) and Gaussian interfering multiple access channels (IMAC), respectively. We focus on regimes of low inter-cell interference, where single-cell transmission with power control and treating inter-cell interference as noise (mc-TIN) is GDoF optimal. Recent works have identified two relevant regimes in this context: one in which the GDoF region achieved through mc-TIN for both the IBC and IMAC is a convex polyhedron without the need for time-sharing (mc-CTIN regime), and a smaller (sub)regime where mc-TIN is GDoF optimal for both the IBC and IMAC (mc-TIN regime). In this work, we extend the mc-TIN framework to cellular scenarios where channel state information at the transmitters (CSIT) is limited to finite precision. We show that in this case, the GDoF optimality of mc-TIN extends to the entire mc-CTIN regime, where GDoF benefits due to interference alignment (IA) are lost. Our result constitutes yet another successful application of robust outer bounds based on the aligned images (AI) approach.

中文翻译：

---

有限精度 CSIT 下小区间干扰作为噪声处理的最优性

在这项工作中，我们研究了下行链路和上行链路蜂窝网络的广义自由度 (GDoF)，分别建模为高斯干扰广播信道 (IBC) 和高斯干扰多址信道 (IMAC)。我们专注于低小区间干扰的机制，其中具有功率控制和将小区间干扰视为噪声 (mc-TIN) 的单小区传输是 GDoF 最佳的。最近的工作已经确定了在这方面的两个相关机制：一个是通过 mc-TIN 为 IBC 和 IMAC 实现的 GDoF 区域是一个不需要分时的凸多面体（mc-CTIN 机制），以及一个较小的（子）体制，其中 mc-TIN 是 IBC 和 IMAC（mc-TIN 体制）的 GDoF 最佳。在这项工作中，我们将 mc-TIN 框架扩展到蜂窝场景，其中发射机的信道状态信息 (CSIT) 仅限于有限精度。我们表明，在这种情况下，mc-TIN 的 GDoF 最优性扩展到整个 mc-CTIN 机制，其中由于干扰对齐 (IA) 的 GDoF 优势丢失。我们的结果构成了基于对齐图像 (AI) 方法的鲁棒外部边界的又一次成功应用。