Adaptive Machine Learning Algorithms for mmWave Communications in Beyond 5G and 6G Systems

利用预算约束的单抗建立了混合频段RF/VLC优化问题，并对不同单抗方案进行了比较。在此基础上，提出了最大数据速率和最小能耗的无人机机载RIS轨迹规划算法。此外，我们将先进的MAB技术（PHE和MOTS）应用于同一问题，并取得了优异的性能结果。为了最大限度地提高bs -用户NLOS联动数据速率和最小化波束训练时间，在RIS中继探测中实现了双目标分组。在多RIS多用户场景中，毫米波RIS-用户关联问题被认为是在保持部署的RIS板之间的负载平衡的同时最大化用户可实现的数据速率。为了解决拟定的强盗博弈问题，并比较它们的性能，提出了三种来自上置信度界（UCB）家族的带有武器负载平衡的集中式MP-MAB算法，即UCB1-LB、Kullback-Leibler UCB- lb （KLUCB-LB）和极大极小最优随机策略- lb （MOSS-LB）。

Hybrid Band RF/VLC optimization problem was formulated using budget constrained MABs and comparsion between different MAB solutions were conducted. Also, we proposed algorithm for UAV mounted RIS trajectory planning that maximizes the data rate and minimizes UAV energy consumption. Besides, we applied sophisticated MAB techniques (PHE and MOTS) to the same problem with superior performance outcome.Dual objective bandits were implemented to RIS relay probing to maximize the BS-user NLOS linkage data rate and minimize the beaform training time.The problem of mmWave RIS-user association in muliple RIS multi-user scenarios is considered to maximize users’ achievable data rates while maintaining load balance among the deployed RIS boards.Three centralized MP-MAB algorithms with arms’ load alancing, coming from the family of upper confidence bound (UCB), namely UCB1-LB, Kullback-Leibler UCB-LB (KLUCB-LB), and minimax optimal stochastic strategy-LB (MOSS-LB), are proposed to address the formulated bandit game and to compare their performance.