Adaptive AI-enhanced computation offloading with machine learning for QoE optimization and energy-efficient mobile edge systems
Abstract Mobile Edge Computing (MEC) systems face critical challenges in optimizing computation offloading decisions while maintaining quality of experience (QoE) and energy efficiency, particularly in dynamic multi-user environments. This paper introduces a novel Adaptive AI-enhanced offloading (AA...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-05-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-00409-4 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Mobile Edge Computing (MEC) systems face critical challenges in optimizing computation offloading decisions while maintaining quality of experience (QoE) and energy efficiency, particularly in dynamic multi-user environments. This paper introduces a novel Adaptive AI-enhanced offloading (AAEO) framework that uniquely integrates three complementary AI approaches: deep reinforcement learning for real-time decision-making, evolutionary algorithms for global optimization, and federated learning for distributed knowledge sharing. The key innovation lies in our hybrid architecture’s ability to dynamically adjust offloading strategies based on real-time network conditions, user mobility patterns, and application requirements, addressing limitations of existing single-algorithm solutions. Through extensive MATLAB simulations with 50–200 mobile users and 4–10 edge servers, our framework demonstrates superior performance compared to state-of-the-art methods. The AAEO framework achieves up to a 35% improvement in QoE and a 40% reduction in energy consumption, while maintaining stable task completion times with only a 12% increase under maximum user load. The system’s security analysis yields a 98% threat detection rate, with response times under 100 ms. Meanwhile, reliability metrics indicate a 99.8% task completion rate and a mean time to failure of 1,200 h. These results validate the proposed hybrid AI approach’s effectiveness in addressing the complex challenges of next-generation MEC systems, particularly in heterogeneous environments requiring real-time adaptation. |
|---|---|
| ISSN: | 2045-2322 |