SDF FuzzIA: A Fuzzy-Based, AI Support System for Decision-Making Frameworks

Abstract

This paper presents an enhanced version of SDF-FuzzIA, a hybrid and modular system designed to support decision-making pro- cesses in geo-thematic domains. Built upon the Sustainability Decision Framework, the system integrates fuzzy ontologies, fuzzy rule-based sys- tems, and a large language model equipped with Retrieval-Augmented Generation capabilities. The proposed architecture aims to deliver accu- rate, explainable, and semantically grounded answers to complex queries involving geopolitical, economic, and environmental indicators in align- ment with the United Nations 2030 Agenda. A novel contribution of this work lies in the adoption of a Multi-Criteria Group Decision-Making methodology to support the construction of fuzzy ontologies, enabling the collaborative selection of membership functions and reducing sub- jectivity in knowledge modeling. Preliminary experiments show promis- ing results in replicating expert reasoning, while highlighting challenges related to scalability, consistency, and automation. This approach lays the groundwork for developing intelligent systems capable of transparent and context-aware decision support in data-rich and uncertainty-prone environments.