A Hybrid Framework of Large Language Models and Retrieval-Augmented Generation for Emergency Command Decision-Making
Abstract
To address the challenges of diverse information sources, rapidly changing situations, and stringent decision-making deadlines in emergency response tasks, this paper proposes a command decision support framework that integrates large language models (LLMs) with retrieval-augmented generation (RAG). The framework consists of six layers: data resources, knowledge enhancement, model capability, application support, security control, and human-computer interaction. It supports tasks such as situation summarization, emergency plan matching, scheme drafting, and after-action evaluation. Based on public emergency event datasets, 90 standardized task samples are constructed across three typical emergency scenarios. Comparative experiments, ablation studies, and parameter sensitivity analysis are then conducted. The results show that, compared with the traditional approach, LLM+RAG improves the comprehensive score by 31.7%. Compared with the pure LLM approach, it increases the Top-3 emergency plan matching accuracy from 68.9% (62/90) to 88.9% (80/90) and reduces the hallucination rate from 14.8% to 4.9%. This study clarifies the theoretical boundaries and division of labor among LLM generation, RAG knowledge enhancement, human-in-the-loop review, and security control. It constructs an evidence-constrained closed-loop generation mechanism, which effectively improves the credibility, traceability and controllability of emergency command decision support.
Read the paper
Related studies
- Knowledge Management Perspective of Generative Artificial Intelligence
- Navigating Generative Artificial Intelligence Promises and Perils for Knowledge and Creative Work
- Characterizing the evolving research on enterprise content management
- Large language models and multimodal foundation models for precision oncology
- From Generative AI to Extended Reality: Multidisciplinary Perspectives on the Challenges, Opportunities, and Future of Educational Computing
- Generative Artificial Intelligence: A New Engine for Advancing Environmental Science and Engineering
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.