Uncertainty-Aware Edge Data Caching with Robust Optimization and Approximate Linear Models

View at publisher Summarise with VIRAT
Authors: P Srinivasarao, Vamsi Kumar Arava, V. Margasahachary, J. Syam Babu, Ch. Mandakini, Swapna Raghunath
Cited by: 0
References: 19

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Edge computing has become essential for supporting low-latency applications by storing popular data closer to users. However, conventional edge data caching techniques often assume predictable demand patterns and overlook real-world uncertainties such as user mobility, abrupt shifts in content popularity, and edge server failures. These assumptions limit their practical effectiveness, especially in dynamic environments where user behavior and system conditions fluctuate unpredictably. In this work, we present an uncertainty-aware edge data caching framework that explicitly accounts for both data popularity variations and potential failures in edge servers. The proposed methodology formulates the caching problem as a two-stage robust optimization task, ensuring resilient performance even under volatile network conditions. To address computational challenges in large-scale scenarios, we further introduce an approximate solution, uEDC-L, based on linear decision rules, significantly reducing overhead while maintaining strong performance. Extensive experimental evaluations demonstrate that our methods outperform established caching strategies, including MagNet and ARO. Notably, the optimal algorithm uEDC-B reduces data retrieval latency by up to 68.4% and lowers storage costs by over 50%, compared to baseline approaches. The approximate method uEDC-L achieves comparable performance with much faster computation, making it a practical choice for real-world deployments. The results confirm that incorporating uncertainty handling into edge caching design can significantly improve both efficiency and system robustness. These contributions advance the reliability of mobile edge computing, especially for applications demanding high data availability under dynamic conditions.

Fields of Study & Subjects

Caching and Content Delivery IoT and Edge/Fog Computing Age of Information Optimization

Publication details

Journal / source
International Conference on Communication, Computing and Electronics Systems (ICCCES)
Pages
1542–1547
Publication date
21 January 2026
Language
EN