Deep Reinforcement Learning-Based Sector Carrier Assignment in CloudRAN
Bibliographic record
Abstract
In the domain of evolving wireless communication technologies such as 5G and beyond, efficient resource allocation is paramount for ensuring service availability and optimal performance. Cloud Radio Access Network (CloudRAN) management offers a dynamic solution to this challenge by integrating cloud-based technologies to optimize resource utilization and enhance scalability. Within CloudRAN, the assignment of Sector Carriers (SCs) plays a critical role in wireless communication systems' operation (e.g., frequency bands), emphasizing the need for efficient data routing to optimal servers. This paper introduces a Deep Reinforcement Learning (DRL) based solution for SC assignment on available server resources, addressing the complex task of balancing conflicting SC’ objectives such as mobility and resiliency. By employing DRL integrated with Graph Neural Networks (GNN), the proposed solution can model complex pairwise relationships between SCs and server resources. The DRL-based solution is trained offline, enabling rapid proposal of SC assignment plans while accommodating diverse requirements such as resiliency, mobility, and resource utilization. Furthermore, the proposed solution dynamically adapts to environmental changes through a feedback loop mechanism, fostering continuous learning. We evaluate the performance of the proposed solution in terms of training convergence, server's capacity violations, and inferencing time. The proposed DRL-GNN solution reduces the number of steps by up to 99% and inference time by over 98% compared to random and Genetic Algorithm (GA) methods. At 1000 SCs, it finds optimal assignments in 9 steps and 2 seconds, while random and GA require over 1000 steps and up to 40 minutes. Additionally, DRL-GNN achieves a 100% success rate, significantly outperforming random (31%) and GA (27%), demonstrating its efficiency and adaptability in optimizing SC assignment
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".