Reasoning about Medical Triage Optimization with Logic Programming
We present a logic programming framework that orchestrates multiple variants of an optimization problem and reasons about their results to support high-stakes medical decision-making. The logic programming layer coordinates the construction and evaluation of multiple optimization formulations, translating solutions into logical facts that support further symbolic reasoning and ensure efficient resource allocation -- specifically targeting the "right patient, right platform, right escort, right time, right destination" principle. This capability is integrated into GuardianTwin, a decision support system for Forward Medical Evacuation (MEDEVAC), where rapid and explainable resource allocation is critical. Through a series of experiments, our framework demonstrates an average reduction in casualties by 35.75% compared to standard baselines. Additionally, we explore how users engage with the system via an intuitive interface that delivers explainable insights, ultimately enhancing decision-making in critical situations. This work demonstrates how logic programming can serve as a foundation for modular, interpretable, and operationally effective optimization in mission-critical domains.