A novel handover scheme for millimeter wave network: An approach of integrating reinforcement learning and optimization

• DOI: 10.1016/j.dcan.2023.08.002
• Publication Date: 2023-08-29T16:41:48Z
• Authors: Ruiyu Wang; Yao Sun; Chao Zhang; Bowen Yang; Muhammad Imran; Lei Zhang
• Keywords: Handover; Reinforcement learning; Optimization theory; MmWave; Information technology; T58.5-58.64; Cellular network

The millimeter-Wave (mmWave) communication with the advantages of abundant bandwidth and immunity to interference has been deemed a promising technology greatly improve network capacity. However, due such characteristics mmWave, as short transmission distance, high sensitivity blockage, large propagation path loss, handover issues (including trigger condition target beam selection) become much complicated. In this paper, we design novel scheme optimize overall system throughput well total delay while guaranteeing Quality Service (QoS) each User Equipment (UE). Specifically, proposed called O-MAPPO integrates Reinforcement Learning (RL) algorithm optimization theory. RL known Multi-Agent Proximal Policy Optimization (MAPPO) plays role in determining conditions. Further, propose an problem conjunction MAPPO select base station. aim is evaluate performance QoS UE after decision made. numerical results show our method are slightly worse than that exhaustive search but better using another typical Deep Deterministic Gradient (DDPG).