Resolving the Scalability Explainability – Adaptability Trilemma in Hybrid Optimization Systems for Disaster Management

Abstract

Hybrid optimization systems have become increasingly critical in disaster management, yet their practical deployment is constrained by a persistent scalability–explainability–adaptability trilemma. This study investigates strategies to resolve this challenge by designing, simulating, and evaluating three hybrid system models across multiple disaster scenarios. Model A prioritized scalability, Model B focused on explainability, and Model C implemented a balanced architecture integrating core optimization, explainability, and adaptive learning layers. Performance metrics included system throughput, interpretability scores, and adaptability under dynamic conditions. Results indicate that while Model A and Model B excelled in their respective targeted dimensions, only Model C demonstrated superior overall performance, achieving an 18% improvement in explainability over Model A, a 22% increase in scalability compared to Model B, and the highest adaptability across all scenarios. The findings highlight the effectiveness of modular architectures, explainable AI integration, and adaptive learning mechanisms in mitigating the trilemma. The study proposes a three-layer hybrid framework that provides a scalable, interpretable, and responsive decision-support system for disaster management. Implications extend to both theory—bridging multi-objective optimization and explainable AI research—and practice, by supporting evidence-based and transparent disaster response strategies