Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Learning-based collaborative MEC for LLM inference with soft-deadline awareness via transformer-enhanced PPO

Nguyen, Ngoc Hung LU orcid and Landfeldt, Björn LU (2026)
Abstract
This paper investigates collaborative mobile edge computing (MEC) servers for large language model (LLM) inference under soft deadline constraints. In this system, to improve the quality of service, computations are expected to be completed within their deadlines. However, due to dependencies among tasks or subtasks, any missed deadline can lead to catastrophic consequences for the entire request. In this context, this work proposes an extended deadline mechanism with constrained flexibility. The main challenges lie in handling large-scale computations under strict latency constraints while limiting the number of allowable deadline extensions, especially in the presence of task dependencies within each request. To tackle these challenges,... (More)
This paper investigates collaborative mobile edge computing (MEC) servers for large language model (LLM) inference under soft deadline constraints. In this system, to improve the quality of service, computations are expected to be completed within their deadlines. However, due to dependencies among tasks or subtasks, any missed deadline can lead to catastrophic consequences for the entire request. In this context, this work proposes an extended deadline mechanism with constrained flexibility. The main challenges lie in handling large-scale computations under strict latency constraints while limiting the number of allowable deadline extensions, especially in the presence of task dependencies within each request. To tackle these challenges, we develop a transformer-enhanced proximal policy optimization (PPO) framework that enables efficient collaboration among MEC servers. The proposed approach aims to maximize the number of tasks completed within their deadlines while minimizing the use of deadline extensions. By capturing temporal dependencies and cross-server interactions, the transformer improves decision-making for task migration. Simulation results demonstrate that the proposed method significantly outperforms conventional PPO and heuristic-based approaches in terms of task completion rate and overall system efficiency. (Less)
Please use this url to cite or link to this publication:
author
and
organization
publishing date
type
Working paper/Preprint
publication status
published
subject
pages
7 pages
publisher
IEEE - Institute of Electrical and Electronics Engineers Inc.
DOI
10.48550/arXiv.2608.02031
language
English
LU publication?
yes
id
f06873f4-3811-4362-be68-33dc13ad714b
date added to LUP
2026-08-01 10:49:13
date last changed
2026-08-31 16:12:00
@misc{f06873f4-3811-4362-be68-33dc13ad714b,
  abstract     = {{This paper investigates collaborative mobile edge computing (MEC) servers for large language model (LLM) inference under soft deadline constraints. In this system, to improve the quality of service, computations are expected to be completed within their deadlines. However, due to dependencies among tasks or subtasks, any missed deadline can lead to catastrophic consequences for the entire request. In this context, this work proposes an extended deadline mechanism with constrained flexibility. The main challenges lie in handling large-scale computations under strict latency constraints while limiting the number of allowable deadline extensions, especially in the presence of task dependencies within each request. To tackle these challenges, we develop a transformer-enhanced proximal policy optimization (PPO) framework that enables efficient collaboration among MEC servers. The proposed approach aims to maximize the number of tasks completed within their deadlines while minimizing the use of deadline extensions. By capturing temporal dependencies and cross-server interactions, the transformer improves decision-making for task migration. Simulation results demonstrate that the proposed method significantly outperforms conventional PPO and heuristic-based approaches in terms of task completion rate and overall system efficiency.}},
  author       = {{Nguyen, Ngoc Hung and Landfeldt, Björn}},
  language     = {{eng}},
  month        = {{08}},
  note         = {{Preprint}},
  publisher    = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}},
  title        = {{Learning-based collaborative MEC for LLM inference with soft-deadline awareness via transformer-enhanced PPO}},
  url          = {{http://dx.doi.org/10.48550/arXiv.2608.02031}},
  doi          = {{10.48550/arXiv.2608.02031}},
  year         = {{2026}},
}