This study investigates eco-environmental zoning-based regulation in China, building a high-quality knowledge graph and multi-agent collaboration to explore intelligent management models.
Eco-environmental zoning-based regulation protects ecological functions and improves environmental quality through differentiated and precise zoning measures. The eco-environmental admittance list integrates control requirements for water, air, soil, ecology, and resource use, covering prohibited, restricted, and permitted access as well as guiding requirements, thus guiding industrial layout and project implementation. However, such lists are voluminous and complex, and their integrated application and analysis remain inadequate. To address this, this study investigated admittance lists in representative regions of China. Through knowledge categorization, it defined seven entity types and eight semantic relationship types, and established an ontology framework for a knowledge graph of eco-environmental zoning-based regulation. It then constructed a knowledge graph with 35,000 entities and 120,000 relationships, and developed an intelligent decision support system integrating graph retrieval-augmented generation (Graph-RAG) with multi-agent collaboration. The results show that knowledge graph construction and multi-agent collaboration overcome technical bottlenecks in traditional admittance list application, improve analysis and judgment in complex tasks, and enhance the precision and efficiency of zoning-based regulation. This study provides a practical reference and a technical approach for the digital and intelligent transformation of eco-environmental management.
Yang Ye, Ph.D., mainly conducts research on environmental impact assessment theory and methods, environmental data management, and applications.