As the popularity of mobile devices and services continues to increase, the demand for more efficient and effective algorithms to manage network utility is increased. Here, we aim to optimize Network Bandwidth Utilization (NBU) by managing user equipments (UEs) and base stations (BSs) association in 5G ORAN architecture. We consider the network total throughput as an objective function and try to find the suitable Cell Individual Offset (CIO) to trigger UE handover with the A3 event such that the network throughput is optimal. We formulate this problem into a Quadratic Unconstrained Binary Optimization (QUBO) problem such that we can explore the power of Quantum Annealing (QA), which is a promising approach to finding the optimal solution for the NP-Hard problem. We also analyze the penalty coefficients setting problem which impacts the outcome of QA significantly. We test our QUBO formulation on a dataset of 7 BSs and 150 UEs with Fujitsu’s third generation Digital Annealer (DA3). Using our QUBO formulation with DA3 solver with fine-tuning, we can improve the throughput with 60% hit rate on this dataset.

随着移动设备和服务的普及，对更高效、更有效的算法来管理网络效用的需求也在增加。在这里，我们的目标是通过管理 5G ORAN 架构中的用户设备 (UE) 和基站 (BS) 关联来优化网络带宽利用率 (NBU)。我们将网络总吞吐量视为目标函数，并尝试找到合适的小区单独偏移（CIO）来触发A3事件的UE切换，从而使网络吞吐量达到最佳。我们将此问题转化为二次无约束二元优化 (QUBO) 问题，以便我们可以探索量子退火 (QA) 的强大功能，这是寻找 NP-Hard 问题最优解的一种很有前途的方法。我们还分析了对 QA 结果产生显着影响的惩罚系数设置问题。我们使用富士通第三代数字退火器 (DA3) 在 7 个 BS 和 150 个 UE 的数据集上测试我们的 QUBO 公式。使用我们的 QUBO 公式和 DA3 求解器并进行微调，我们可以将该数据集上的命中率提高 60%。