面向智能体互联网的AIGC任务分布式自适应智能卸载策略OA
Distributed adaptive intelligent offloading strategies for AIGC tasks in the Internet of agents
针对智能体互联网环境下人工智能生成内容(AIGC)任务的高算力需求与动态服务质量特性,提出了一种基于PPO的智能计算卸载策略.首先,构建包含卸载位置决策与生成质量等级选择的二维联合动作空间,以适应AIGC任务的可伸缩特性;其次,设计融合用户偏好模式的用户服务体验质量奖励函数.进一步地,针对多维状态空间量纲差异导致的训练不稳定,引入状态与奖励双重在线归一化机制,并结合动态学习率调度策略,提升算法对复杂环境的特征提取效率.仿真结果表明,所提算法在收敛速度与稳定性上显著优于基准算法.
The Internet of agents environment imposes high computational demands and requires dynamic quality of ser-vice(QoS)for artificial intelligence generated content(AIGC)tasks,an intelligent computational offloading strategy based on PPO was proposed.Firstly,a two-dimensional joint action space encompassing offloading location decisions and generation quality level selection was constructed to accommodate the scalable characteristics of AIGC tasks.Sec-ondly,a quality of experience reward function integrating user preference modes was designed.Furthermore,to address training instability caused by dimensional discrepancies in the multidimensional state space,a dual online normalization mechanism for states and rewards was introduced,which was combined with a dynamic learning rate scheduling strategy to enhance the algorithm's feature extraction efficiency in complex environments.Simulation results demonstrate that the proposed algorithm significantly outperforms baseline algorithms in terms of convergence speed and stability.
袁晓铭;张馨灵;邓庆绪;李长乐;王嘉诚;策力木格
东北大学秦皇岛分校河北省海洋感知网络与数据处理重点实验室,河北 秦皇岛 066004||湖北大学大数据智能分析与行业应用湖北省重点实验室,湖北 武汉 430062东北大学秦皇岛分校河北省海洋感知网络与数据处理重点实验室,河北 秦皇岛 066004东北大学秦皇岛分校河北省海洋感知网络与数据处理重点实验室,河北 秦皇岛 066004西安电子科技大学空天地一体化综合业务网全国重点实验室,陕西 西安 710071新加坡南洋理工大学计算机与数据科学学院,新加坡 308232电气通信大学信息理工学研究科,东京都 调布 182-8585
信息技术与安全科学
智能体互联网边缘侧大模型移动边缘计算深度强化学习人工智能生成内容
Internet of agentsedge-based large modelmobile edge computingdeep reinforcement learningAIGC
《通信学报》 2026 (7)
65-79,15
国家自然科学基金资助项目(No.62371116)河北省自然科学基金优秀青年科学基金资助项目(No.F2026501042)大数据智能分析与行业应用湖北省重点实验室开放基金(湖北大学)资助项目(No.2025BDIAA03)河北省教育厅在读研究生创新能力培养基金资助项目(No.CXZZSS2026175) The National Natural Science Foundation of China(No.62371116),Hebei Natural Science Fund for Excellent Young Scientists(No.F2026501042),Open Fund of Hubei Key Laboratory of Big Data Intelligent Analysis and Industrial Applica-tions(Hubei University)(No.2025BDIAA03),Hebei Provincial Department of Education Funding Program for Cultivating Innova-tive Ability of Postgraduate Students(No.CXZZSS2026175)
评论