基于多智能体深度强化学习的智慧医疗网络计算卸载方法  

Multi-agent Deep Reinforcement Learning Based Edge Computing Offloading Method in Smart Healthcare Network

作　　者：方馨 秦子健 高新平 闻煜超 苏新[1] FANG Xin;QIN Zijian;GAO Xinping;WEN Yuchao;SU Xin(The College of Information Science and Engineering,Hohai University,Changzhou 213022,China)

机构地区：[1]河海大学信息科学与工程学院,江苏常州213022

出　　处：《移动通信》2024年第6期86-90,共5页Mobile Communications

基　　金：国家自然科学基金项目(62371181)“面向海洋移动边缘计算的任务卸载与隐私保护关键技术研究”;常州市政策引导类计划“常州文化遗产智慧复原与沉浸式交互文化产业应用示范”(国际科技合作/港澳台科技合作CZ20230029)。

摘　　要：6G和移动边缘计算技术的重大突破推动了智慧医疗服务的繁荣发展。为了满足医疗服务的低时延、高可靠需求,提出了一种基于多智能体深度强化学习的计算卸载方法。考虑时延和能耗,建立混合整数非线性规划任务卸载问题。通过传统优化算法求解资源分配问题,并通过多智能体深度强化学习算法求解卸载决策问题。仿真结果表明,与现有的几种方法相比,所提算法能够在智慧医疗网络动态变化环境中进行实时任务卸载。The significant breakthroughs in 6G and mobile edge computing technologies have driven the flourishing development of smart healthcare services.In order to meet the requirements of low latency and high reliability in healthcare services,a computational offloading method based on multi-agent deep reinforcement learning is proposed.Considering the latency and energy consumption,a mixed integer nonlinear programming(MINLP)task offoading problem is formulated.The resource allocation problem is solved by a traditional optimization algorithm and the offloading decision problem is solved by a multi-agent deep reinforcement learning algorithm.Simulation results show that the proposed algorithm can perform real-time task offloading in dynamically changing environments of smart healthcare networks compared with several existing approaches.

关 键 词：智慧医疗网络 深度强化学习 移动边缘计算 资源分配 计算卸载 

分 类 号：TN929.5[电子电信—通信与信息系统]