Scholarlink

The increasing reliance on live video streaming technology has heightened the demand for adaptive and cognitive service infrastructures. However, cloud-based models are faced with enormous challenges of high latency, bandwidth limitations, and dynamic workloads, which impinge the Quality of Service (QoS) of multimedia applications. To address these issues, this paper proposes an adaptive agentic AI-based multimedia architecture for Cloud–Edge service optimization, designed to dynamically manage resource selection and allocation in heterogeneous environments. Our architecture leverages agentic AI to adaptively assign service requests (e.g., educational content) to distributed cloud and edge nodes based on current latency and load metrics. The system models a Mobile Edge Computing (MEC) simulation that integrates lightweight autonomous decision-making agents capable of selecting optimal service nodes. The architecture ensures improved Quality of Experience (QoE), low latency delivery, and efficient edge resource utilization. Experimental evaluation, supported by visual analytics, demonstrates the framework's potential to meet service objectives in varying network conditions, providing a strong foundation for integrating more advanced agentic AI models in decentralized multimedia service environments. This early-stage ‘agentic AI’ approach emphasizes the scheduler’s autonomy and real-time adaptability via a lightweight heuristic, providing immediate performance benefits at the edge while laying the groundwork for integrating more advanced learning-based intelligence in future deployments.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Agentic AI-Based Architecture for E-learning Multimedia Service Selection in Cloud–Edge Environments

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Asma Saighi ,
  • Soumia Zertal,
  • Ahlem Nasri,
  • Zakaria Laboudi,
  • Abdellah Kouzou,
  • Mohamed Deriche,
  • Makhlouf Derdour

摘要

The increasing reliance on live video streaming technology has heightened the demand for adaptive and cognitive service infrastructures. However, cloud-based models are faced with enormous challenges of high latency, bandwidth limitations, and dynamic workloads, which impinge the Quality of Service (QoS) of multimedia applications. To address these issues, this paper proposes an adaptive agentic AI-based multimedia architecture for Cloud–Edge service optimization, designed to dynamically manage resource selection and allocation in heterogeneous environments. Our architecture leverages agentic AI to adaptively assign service requests (e.g., educational content) to distributed cloud and edge nodes based on current latency and load metrics. The system models a Mobile Edge Computing (MEC) simulation that integrates lightweight autonomous decision-making agents capable of selecting optimal service nodes. The architecture ensures improved Quality of Experience (QoE), low latency delivery, and efficient edge resource utilization. Experimental evaluation, supported by visual analytics, demonstrates the framework's potential to meet service objectives in varying network conditions, providing a strong foundation for integrating more advanced agentic AI models in decentralized multimedia service environments. This early-stage ‘agentic AI’ approach emphasizes the scheduler’s autonomy and real-time adaptability via a lightweight heuristic, providing immediate performance benefits at the edge while laying the groundwork for integrating more advanced learning-based intelligence in future deployments.