重庆邮电大学学报（自然科学版）2025，Vol.37Issue(2)：196-203,8.DOI:10.3979/j.issn.1673-825X.202407030165
车联网预感知服务请求及服务迁移机制
Anticipatory service request and service migration mechanism in internet of vehicles
摘要
Abstract
With the rapid development of real-time vehicular applications,the data traffic and content requests in the inter-net of vehicles have surged dramatically.To ensure that vehicular users can obtain real-time and stable services while opti-mizing system costs,a presentiment service request mechanism for internet of vehicles scenarios is designed.This mecha-nism can make presentiment decisions based on environmental information,thereby reducing the costs associated with the migration process.A speed-based service migration optimization algorithm is proposed,which dynamically adjusts migration decisions according to the real-time speed of vehicles to ensure the efficiency and stability of the migration process.Consid-ering the uncertainty in the interactions between vehicular users,roadside units,and base stations,a presentiment multi-a-gent deterministic policy gradient algorithm is proposed to optimize migration strategies.By interacting with the environ-ment,agents gradually adjust their strategies to maximize long-term cumulative rewards.Experimental results demonstrate that the proposed algorithms effectively reduce system costs and significantly enhance system performance.关键词
车联网/服务迁移/多智能体深度强化学习Key words
internet of vehicle/service migration/multi-agent deep reinforcement learning分类
电子信息工程引用本文复制引用
陈冰旖,左琳立,李云..车联网预感知服务请求及服务迁移机制[J].重庆邮电大学学报（自然科学版）,2025,37(2):196-203,8.基金项目
重庆市自然科学基金创新发展联合基金项目(CSTB2022NSCQ-LZX0055) Chongqing Natural Science Foundation Innovation Development Joint Fund(CSTB2022NSCQ-LZX0055) （CSTB2022NSCQ-LZX0055）
