Online Edge-Agent Bitrate Control Under Memory Constraints: A Reinforcement Learning Approach
International Journal of Business and Social Science Studies · 19.09.2026 · 6 (2) · 274-286
DOI: 10.66795/ijbsss78 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Business and Social Science Studies, online erschienen am 19.09.2026. Verfasst von Daniel A. Phillips und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The exponential growth of multimedia traffic has placed unprecedented pressure on global telecommunications infrastructure, necessitating advanced traffic management and video delivery paradigms. Edge computing has emerged as a transformative architecture, bringing computational resources and caching capabilities closer to end-users to reduce latency and alleviate backhaul congestion. Within this paradigm, adaptive bitrate streaming managed by intelligent edge agents has shown significant promise in enhancing the Quality of Experience for users. However, edge nodes inherently suffer from stringent resource limitations, particularly regarding memory and storage, which fundamentally restricts the state space and historical data that online learning agents can utilize for optimal decision-making. This paper presents a comprehensive investigation into online edge-agent bitrate control under strict memory constraints. By formulating the bitrate adaptation process as a constrained Markov Decision Process, we introduce a novel memory-aware online reinforcement learning framework that explicitly incorporates memory utilization penalties into the reward function. Through extensive theoretical analysis and rigorous system-level simulations, we evaluate the performance of the proposed mechanism against conventional heuristics and unconstrained learning algorithms. The methodology dynamically balances the trade-off between maximizing video quality, minimizing rebuffering events, and adhering to strict memory bounds. Our results demonstrate that the proposed memory-aware agent achieves superior bandwidth utilization and maintains a high Quality of Experience while reducing memory overhead by a significant margin compared to traditional approaches.
Zitieren
APA
Phillips, D. A., & Ward, H. (2026). Online Edge-Agent Bitrate Control Under Memory Constraints: A Reinforcement Learning Approach. International Journal of Business and Social Science Studies, 6(2), 274-286. https://doi.org/10.66795/ijbsss78
Vancouver
Phillips DA, Ward H. Online Edge-Agent Bitrate Control Under Memory Constraints: A Reinforcement Learning Approach. International Journal of Business and Social Science Studies. 2026;6(2):274-286. doi:10.66795/ijbsss78
BibTeX
@article{Phillips2026,
author = {Phillips, Daniel A. and Ward, Harry},
title = {Online Edge-Agent Bitrate Control Under Memory Constraints: A Reinforcement Learning Approach},
journal = {International Journal of Business and Social Science Studies},
year = {2026},
volume = {6},
number = {2},
pages = {274--286},
doi = {10.66795/ijbsss78}
}
EUVIMED