Hierarchical Reinforcement Learning with Dynamic Meta Agent for Adaptive Cut Selection in Integer Programming with Applications to Sensor Network Design
Bibliographic record
Abstract
Designing optimal sensor networks for industrial processes presents significant combinatorial challenges. These are often formulated as large-scale integer programming (IP) problems. Efficiently solving these IPs, particularly through cutting plane selection, remains a critical bottleneck, due to its complexity. Cutting plane selection is a fundamental technique in many IP solvers. It involves iteratively adding valid linear inequalities (cuts) to the problem formulation to eliminate noninteger solutions without removing any feasible integer solutions. This work proposes a hierarchical reinforcement learning (HRL) framework with dynamic agent allocation to enhance adaptive cutting plane selection for the solution of these problems. Using a two-level architecture powered by the Proximal Policy Optimization (PPO) backbone, the framework features a meta-agent that selects cutting strategies and allocates lower-level agents to refine LP relaxations. We evaluate the approach on challenging sensor network design problems and synthetic IP instances, comparing it to single-agent RL and traditional heuristics. Results show that HRL achieves higher cumulative rewards, improved success rates, and faster convergence, achieving over 25% and 60% time savings for Gomory and cover cuts, respectively, on these large instances, demonstrating its scalability and effectiveness in optimizing cut selection for real-world applications.
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 | 0.001 | 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".