Joint Optimization of Resource Allocation and Topology Formation for Hierarchical Federated Learning in Smart Grids
Bibliographic record
Abstract
Hierarchical Federated Learning (HFL) presents a promising approach for addressing communication challenges inherent in traditional FL methodologies, particularly critical in resource-constrained environments such as Smart Grids (SG). In this paper, we delve into the potential of HFL tailored to the clustered structure of typical SG use-cases such as load forecasting, particularly in scenarios characterized by high communication time and energy constraints. We propose a novel optimization model that jointly tackles topology formation and resource allocation in HFL. Unlike prior works, our model handles an unknown number of layers, allowing for flexible adaptation to different grid configurations. We solve this non-convex optimization model by an approximation method that separates clustering from resource allocation. Clustering is addressed using complete-linkage clustering and Voronoi diagrams, while resource allocation is tackled through approximations and convexification of the problem. Furthermore, we introduce a Deep Q-Network (DQN) based reinforcement learning approach that addresses the joint optimization problem in near real-time. Simulation results demonstrate that the hierarchical architecture with efficient resource allocation significantly reduces energy consumption in smart grid sensor nodes compared to traditional FL, while meeting response time requirements. In addition, our DQN solution achieves near-optimal performance, making it particularly appropriate for large-scale grids.
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 | 0.004 |
| 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 | 0.002 |
| 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".