Reinforcement Learning for Server-Aware Offloading in Multi-Tier Multi-Instance Computing Architecture
Open Access
- 17 January 2026
- journal article
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Intelligent Systems and Technology
- Vol. 17 (1) , 1-25
Abstract
Task offloading in distributed computing involves complex tradeoffs among delay, scalability, cost, and resource utilization. Cloud platforms face long communication delays, while edge nodes have constrained capacity. Static, rule-based schedulers cannot adapt to fluctuating loads or per-instance heterogeneity. Similarly, Reinforcement Learning (RL) schemes typically address only a single layer or assume homogeneous servers, overlooking the hierarchical and multi-instance nature of deployments. To address these challenges, we introduce a server-aware Proximal Policy Optimization (PPO) framework that performs fine-grained offloading across a three-tier (Edge, Regional, Cloud) multi-instance architecture. We formulate offloading as a Markov Decision Process whose state vector includes per-instance delay, CPU/memory utilization, network congestion, cost, and energy metrics. The PPO agent learns to offload tasks to the best server in real time. Our developed RegionalEdgeSimPy simulation shows that PPO agent makes optimal offloading choices for over 90% of tasks, keeping each server near; however, below a 70% utilization. This optimized decision making drives up to 66.9% delay reduction, 78.6% energy savings, and 47.8% cost reductions relative to cloud-only and edge-only baselines.Keywords
This publication has 45 references indexed in Scilit:
- Solving Scalable Multiagent Routing Problems With Reinforcement LearningIEEE Transactions on Neural Networks and Learning Systems, 2025
- Task offloading delay minimization in vehicular edge computing based on vehicle trajectory predictionDigital Communications and Networks, 2025
- Social big data management through collaborative mobile, regional, and cloud computingPeerJ Computer Science, 2025
- Regional computing approach for educational big dataScientific Reports, 2025
- Multiple Edge Data Integrity Verification With Multi-Vendors and Multi-Servers in Mobile Edge ComputingIEEE Transactions on Mobile Computing, 2025
- Big data applications: overview, challenges and futureArtificial Intelligence Review, 2024
- Improved particle swarm optimization with reverse learning and neighbor adjustment for space surveillance network task schedulingSwarm and Evolutionary Computation, 2024
- Why and How Lasagna Works: A New Design of Air-Ground Integrated InfrastructureIEEE Network, 2024
- Multiobjective Offloading Optimization in Fog Computing Using Deep Reinforcement LearningJournal of Computer Networks and Communications, 2024
- Deep Reinforcement Learning for Task Offloading in Mobile Edge Computing SystemsIEEE Transactions on Mobile Computing, 2020