Reinforcement learning-based dynamic load balancing in edge computing networks
Bibliographic record
Abstract
Edge computing (EC) has emerged as a paradigm aimed at reducing data transmission latency by bringing computing resources closer to users. However, the limited scale and constrained processing power of EC pose challenges in matching the resource availability of larger cloud networks. Load balancing (LB) algorithms play a crucial role in distributing workload among edge servers and minimizing user latency. This paper presents a novel set of distributed LB algorithms that leverage machine learning techniques to overcome the three limitations of our previous LB algorithm, EVBLB : (i) its reliance on static time intervals for execution, (ii) the need for comprehensive information about all server resources and queued requests for neighbor selection, and (iii) the use of a central coordinator to dispatch incoming user requests over edge servers. To offer increased control, custom configuration, and scalability for LB on edge servers, we propose three efficient algorithms: Q-learning (QL), multi-armed bandit (MAB), and gradient bandit (GB) algorithms. The QL algorithm predicts the subsequent execution time of the EVBLB algorithm by incorporating rewards obtained from previous executions, thereby improving performance across various metrics. The MAB and GB algorithms prioritize near-optimal neighbor node servers while considering dynamic changes in request rate, request size, and edge server resources. Through simulations, we evaluate and compare the algorithms in terms of network throughput, average user response time , and a novel LB metric for workload distribution across edge servers.
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 | 0.003 |
| 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) | 0.001 | 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".