Using Reinforcement Learning to Provide Decision Support in Multi-Domain Mass Evacuation Operations
Bibliographic record
Abstract
In this paper, we study a scenario in which a large number of individuals in various levels of medical distress are stranded at a remote location, such as in the Arctic, and must be evacuated. Set within this context, we examine a multi-domain operation in which the evacuation of individuals occurs via one of two ways, either by helicopter or by ship, each with their own capacity constraints. The aim of this research is to determine a decision policy whose objective is to maximize the number of survivors. This is achieved by seeking a policy that throughout the operation effectively coordinates the selection of those individuals to be evacuated via helicopter and those to be evacuated via ship. Our contributions are twofold. First, we formulate the multi-domain mass evacuation operation as a Markov Decision Process. Second, due to the fact that the curse of dimensionality renders exact methods not applicable, we employ an Artificial Intelligence framework, namely, Reinforcement Learning (RL), also known as Approximate Dynamic Programming (ADP) within operations research, to learn a near-optimal policy. Using a value function approximation based on state aggregation, we design an ADP algorithm to learn a policy within the context of a representative planning scenario. We then apply this policy across a range of test scenarios and compare the outcomes to those achieved using non-coordinated benchmark policies. Although our learned policy does not outperform all benchmarks, our results demonstrate how Artificial Intelligence may be used to evaluate candidate policies and provide decision support in multi-domain operations.
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 | 0.002 | 0.009 |
| 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 | 0.002 |
| 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".