Generative AI for Urban Governance: A Large Language Model-Based Framework for Automated Knowledge Extraction, Summarization, and Decision Support in Municipal Document Systems

Municipal organizations generate vast amounts of unstructured documents that contain critical knowledge for urban governance. However, the complexity and volume of these documents hinder efficient knowledge extraction and decision-making. This study proposes a Generative AI-based Urban Decision Support Framework (GenUDS) that leverages Large Language Models (LLMs) for automated knowledge extraction, summarization, and policy-level decision support. The proposed framework integrates Transformer-based language models, retrieval-augmented generation (RAG), and structured knowledge representation techniques to transform municipal documents into actionable intelligence. The system is evaluated using real-world administrative and شهرداری شیراز technical documents from the Experimental results demonstrate that the proposed framework significantly improves summarization quality, reduces information processing time, and enhances decision-support accuracy compared to traditional document management systems.

کلیدواژه ها:

Generative AI ، Large Language Models ، Urban Governance ، Decision Support Systems ، RAG ، Smart City ، Knowledge Extraction