基于学习的下一代智能网络方法

Thesis by Liang Zhang In Partial Fulfillment of the Requirements For the Degree of Doctor of Philosophy King Abdullah University of Science and TechnologyThuwal, Kingdom of Saudi Arabia EXAMINATION COMMITTEE PAGE The thesis of Liang Zhang is approved by the examination committee Committee Chairperson: Prof. Basem ShihadaCommittee Members: Prof. Mohamed-Slim Alouini, Prof. Suhaib Fahmy, Prof. RaduStoleru ©April, 2022Liang ZhangAll Rights Reserved ABSTRACT Learning-Based Approaches for Next-Generation Intelligent NetworksLiang Zhang The next-generation (6G) networks promise to provide extended 5G capabilitieswith enhanced performance at high data rates, low latency, low energy consumption,and rapid adaptation. 6G networks are also expected to support the unprecedentedInternet of Everything (IoE) scenarios with highly diverse requirements.With theemerging applications of autonomous driving, virtual reality, and mobile computing,achieving better performance and fulfilling the diverse requirements of 6G networksare becoming increasingly difficult due to the rapid proliferation of wireless dataand heterogeneous network structures. In this regard, learning-based algorithms arenaturally powerful tools to deal with the numerous data and are expected to impactthe evolution of communication networks. This thesis employed learning-based approaches to enhance the performance andfulfill the diverse requirements of the next-generation intelligent networks under vari-ous network structures. Specifically, we design the trajectory of the unmanned aerialvehicle (UAV) to provide energy-efficient, high data rate, and fair service for the Inter-net of things (IoT) networks by employing on/off-policy reinforcement learning (RL).Thereafter, we applied a deep RL-based approach for heterogeneous traffic offloadingin the space-air-ground integrated network (SAGIN) to cover the co-existing require-ments of ultra-reliable low-latency communication (URLLC) traffic and enhancedmobile broadband (eMBB) traffic. Precise traffic prediction can significantly improvethe performance of 6G networks in terms of intelligent network operations, such aspredictive network configuration control, traffic offloading, and communication re-source allocation.Therefore, we investigate the wireless traffic prediction problem in edge networks by applying a federated meta-learning approach. Lastly, we designan importance-oriented clustering-based high quality of service (QoS) system withsoftware-defined networking (SDN) by adopting unsupervised learning. ACKNOWLEDGEMENTS I would like to thank my supervisor, Prof.Basem Shihada, who supported mefrom various perspectives during my Ph.D. I would also like to thank the committeemembers, Prof. Mohamed-Slim Alouini, Prof.Suhaib Fahmy, and Prof. Radu Stoleru,for their time and patience. I would like to thank my co-authors in Netlab, Dr. WiemAbderrahim, Dr. Abdulkadir Celik, Dr. Shuping Dang, and Dr. Chuanting Zhang,for their support. Finally, I would like to thank my family and friends TABLE OF CONTENTS Examination Committee Page2 Copyright3 4 Abstract Table of Contents7 List of Figures11 List of Tables13 1Introduction 1.1Thesis Statement and Motivation. . . . . . . . . . . . . . . . . . . .141.2Thesis objective and Contributions. . . . . . . . . . . . . . . . . . .161.3Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 2Background and State of the Art20 2.1Overview of Machine Learning in Communication Networks. . . . .202.2Category of Machine Learning Methods. . . . . . . . . . . . . . . .202.2.1Unsupervised Learning in Communication Networks . . . . . . . . .202.2.2Supervised Learning in Communication Networks. . . . . . . . . .212.2.3Reinforcement Learning in Communication Networks. . . . . . . .222.2.4Federated Learning and Meta Learning in Communication Networks:222.3State of the Art . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .232.3.1Energy-Efficient Trajectory Optimization for UAV-Assisted IoT Net-works. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .232.3.2Heterogeneous Traffic Offloading for SAGIN. . . . . . . . . . . . .27 2.3.3Wireless Traffic Prediction at Edge Networks. . . . . . . . . . . .292.3.4Adaptive QoS System with SDN. . . . . . . . . . . . . . . . . . .31 3Energy-Efficient Trajectory Optimization for UAV-Assisted IoT Network343.1System Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .353.1.1Network Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . .353.1.2Energy Harvesting Model. . . . . . . . . . . . . . . . . . . . . . .363.1.3Energy Consumption Models. . . . . . . . . . . . . . . . . . . . .393.1.4Channel Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . .423.1.5Fairness Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . .443.2Problem Statement, Formulation, and Solution . . . . . . . . . . . . .453.2.1Problem Statement: A Multi-Objective Trajectory Design