AI 中文总结
研究开放无线接入网中节能与覆盖/吞吐量xApp的直接冲突，提出双保真感知硬切换仲裁器，通过监控预测与观察效用误差解决冲突，在5G系统级评估中实现接近最优的吞吐量-功率权衡，降低效用遗憾。
AI 中文摘要
开放无线接入网（O-RAN）允许独立开发的xApp通过近实时RAN智能控制器（Near-RT RIC）控制RAN功能。当目标冲突的xApp并发运行时，可能会发出不兼容操作，降低网络性能。本文解决节能（ES）xApp和面向覆盖/吞吐量（CTO）xApp对同一小区下行传输功率设置不同的直接冲突。将冲突解决制定为在线选择两个提议的连续混合，最大化联合考虑吞吐量和功耗的能量感知效用。网络数字孪生（NDT）在实时部署前预测候选操作的效用，但当孪生漂移时，选择最高孪生预测效用变得无效。因此提出双保真感知硬切换仲裁器，使用指数加权移动平均监控预测效用和观察效用之间的误差。误差低于阈值时，仲裁器遵循NDT选择的操作；否则切换到在线学习的最佳先前观察操作。仲裁器轻量级、无需训练且无需最优策略的预言知识。5G系统级评估表明，它在运营商能量优先级上实现了最接近最优的吞吐量-功率权衡，归一化效用遗憾为$0.017 \pm 0.006$，而基于COMIX风格孪生选择器为$0.159 \pm 0.052$。在严重NDT漂移（10 dB）下，效用遗憾从$11.19 \pm 3.58$降至$0.55 \pm 0.25$。这些结果表明在线孪生保真度监控可实现强大的数字孪生辅助xApp冲突解决，同时保持效用感知吞吐量-功率优化。
英文摘要
Open Radio Access Network (O-RAN) allows independently developed xApps to control RAN functions through the Near-Real-Time RAN Intelligent Controller (Near-RT RIC). When xApps with conflicting objectives operate concurrently, they may issue incompatible actions that degrade network performance. This paper addresses a direct conflict in which an energy-saving (ES) xApp and a coverage/throughput-oriented (CTO) xApp request different downlink transmit-power settings for the same cell. We formulate conflict resolution as online selection of a continuous blend of the two proposals, maximizing an energy-aware utility that jointly considers throughput and power consumption. A network digital twin (NDT) predicts this utility for candidate actions before live deployment, but selecting the highest twin-predicted utility becomes ineffective when the twin drifts. We therefore propose a twin-fidelity-aware hard-switching arbiter that monitors the error between predicted and observed utilities using an exponentially weighted moving average. While the error remains below a threshold, the arbiter follows the NDT-selected action; otherwise, it switches to the best previously observed action learned online. The arbiter is lightweight, training-free, and requires no oracle knowledge of the optimal policy. System-level 5G evaluations show that it achieves the closest throughput-power trade-off to the optimum across operator energy priorities, yielding normalized utility regret of $0.017 \pm 0.006$, versus $0.159 \pm 0.052$ for a COMIX-style twin-based selector. Under severe NDT drift (10 dB), it reduces utility regret from $11.19 \pm 3.58$ to $0.55 \pm 0.25$. These results show that online twin-fidelity monitoring enables robust digital-twin-assisted xApp conflict resolution while preserving utility-aware throughput-power optimization.
Comments15 pages, 10 figures. Submitted to IEEE Transactions on Network and Service Management
