Scheduling improves mobile edge AI task completion times
Workflow-Aware Communication and Computation Scheduling for Mobile Edge AI Agents
Networking and Internet Architecture
Summary
Mobile AI agents running on edge devices have to handle complex tasks that can change depending on previous results. The authors look at how to best schedule these tasks and the sharing of wireless and computing resources in real time to reduce delays. They create a model that predicts task progress and uses this to plan resource allocation efficiently. Their simulations show this scheduling method performs better than current approaches in various settings.
What this means in practice
- •For mobile network operators: Optimize computing and wireless resource distribution to speed up task completion for AI services running on edge devices.
- •For iot system integrators: Improve responsiveness of edge AI agents in IoT deployments where task workflows are dynamic and uncertain.
Authors
Xingqiu He, Chaoqun You, Feng Wang, Zihan Chen, Shuai Wang, Jihong Park
Abstract
AI agents are increasingly deployed in mobile edge environments to autonomously handle diverse tasks through multi-step workflows. Unlike conventional dependency-aware tasks with predefined dependency structures, agent workflows are dynamically generated based on intermediate execution results. To minimize the weighted sum of job completion latencies, this paper models partially revealed agent workflows using dynamic workflow graphs and formulates an online scheduling problem that jointly optimizes computation placement, wireless resource block allocation, and edge computing resource allocation. To address the uncertainty of future workflow evolution, we first transform the weighted completion time objective into an equivalent per-slot holding cost and characterize the optimal policy through a Bellman equation. We then construct a tractable look-ahead cost-to-go surrogate based on the revealed workflow and the estimated remaining latencies of active jobs. The proposed scheduling algorithm leverages the surrogate to determine computation placement and uses the resulting operation criticality to allocate wireless and edge computing resources online. Simulation results show that the proposed algorithm outperforms the benchmark schemes across diverse scenarios.