RESEARCH ON MULTI-AGENT LARGE LANGUAGE MODELS (MULTI-AGENT LLMs) FOR AUTOMATIC GEOGRAPHIC MAP GENERATION
Keywords
DOI:
https://doi.org/10.34238/tnu-jst.14758Abstract
Advancement in GeoAI and generative AI is transforming the acquisition, analysis, and visualization of geospatial data. The Autonomous GIS paradigm leverages the natural language understanding and reasoning capabilities of large language models (LLMs) to automate GIS tasks based on natural language requests. However, as tasks become increasingly complex,featuring multiple data layers, multi-step processing, and cartographic constraints,single-agent systems face limitations in planning, error control, and ensuring the spatial accuracy and consistency of the results.This paper proposes a multi-agent LLM-based framework for automated map generation following the workflow: “user intent - data retrieval - spatial analysis - visualization - evaluation”. Experiments conducted using real-world OpenStreetMap data, including buffer creation and spatial join operations, demonstrate that the multi-agent architecture effectively separates functional roles, enhances scalability, and improves the transparency of the map generation process.Downloads
References
K. Janowicz, S. Gao, G. McKenzie, Y. Hu, and B. Bhaduri, “GeoAI: Spatially explicit artificial intelligence techniques for geographic knowledge discovery and beyond,” International Journal of Geographical Information Science, vol. 34, no. 4, pp. 625-636, 2020, doi: 10.1080/13658816.2019.1684500.
Z. Li and H. Ning, “Autonomous GIS: the next-generation AI-powered GIS,” International Journal of Digital Earth, vol. 16, no. 1, pp. 4385-4407, 2023, doi: 10.1080/17538947.2023.2278895.
Z. Li, H. Ning, S. Gao, K. Janowicz, W. Li, S. T. Arundel, Y. Hu, and B. Bhaduri, “GIScience in the era of artificial intelligence: A research agenda towards autonomous GIS,” Annals of GIS, vol. 31, no. 1, pp. 1-18, 2025, doi: 10.1080/19475683.2025.2552161.
T. Brown et al., “Language models are few-shot learners,” Advances in Neural Information Processing Systems, vol. 33, pp. 1877-1901, 2020.
J. Achiam et al., “GPT-4 technical report,” arXiv:2303.08774, 2023, doi: 10.48550/arXiv.2303.08774.
Y. Zhang, C. Wei, S. Wu, Z. He, and W. Yu, “GeoGPT: Understanding and processing geospatial tasks through an autonomous GPT,” arXiv:2307.07930, 2023, doi: 10.48550/arXiv.2307.07930.
E. Unlu, “ChatMap: Large language model interaction with cartographic data,” arXiv:2310.01429, 2023, doi: 10.48550/arXiv.2310.01429.
Q. Lin, R. Hu, H. Li, S. Wu, Y. Li, K. Fang, H. Feng, W. Yu, and X. Song, “ShapefileGPT: A multi-agent large language model framework for automated shapefile processing,” arXiv:2410.12376, 2024, doi: 10.48550/arXiv.2410.12376.
C. Lee, V. Paramanayakam, A. Karatzas, Y. Jian, S. K. Ramchurn, J. Pottinger, S. Madden, S. Kim, L. D. Cruz, and D. Stamoulis, “Multi-Agent Geospatial Copilots for Remote Sensing Workflows,” arXiv:2501.16254, 2025, doi: 10.48550/arXiv.2501.16254.
H. Ning, Z. Li, T. Akinboyewa, and M. Just, “An autonomous GIS agent framework for geospatial data retrieval,” International Journal of Digital Earth, vol. 18, no. 1, 2025, doi: 10.1080/17538947.2025.2458688.
C. Sun, T. Lan, and Z. Wu, “Generative artificial intelligence and its applications in cartography and GIS: an exploratory review,” Journal of Geodesy and Geoinformation Science, vol. 8, no. 1, pp. 1-20, 2025, doi: 10.11947/j.JGGS.2025.0205.
C. Wang, Y. Kang, Z. Gong, P. Zhao, Y. Feng, S. Gao, and Q. Zhou, “CartoAgent: a multimodal large language model-powered multi-agent cartographic framework for map style transfer and evaluation,” International Journal of Geographical Information Science, vol. 39, pp. 1904-1937, 2025, doi: 10.1080/13658816.2025.2507844.
Government of Vietnam, “Resolution No. 40/NQ-CP on the Strategy for the Development of Vietnam’s Surveying and Mapping Sector and the Development of the National Geospatial Data Infrastructure to 2030, with a Vision to 2045,” (in Vietnamese), Mar. 27, 2023. [Online]. Available: https://datafiles.chinhphu.vn/ cpp/files/vbpq/2023/3/40-nq.signed.pdf. [Accessed Feb. 05, 2026].
K. S. Nguyen, T. V. Nguyen, and D.-B. Nguyen, “A systematic review of Artificial Intelligence in Geographic Information Systems,” in International Conference on Advances in Information and Communication Technology, Springer, 2023, pp. 20-31, doi: 10.1007/978-3-031-49529-8_3.
K. S. Nguyen, T. V. Nguyen, H. H. Ngo, and D. B. Nguyen, “Application of Large Language Models in Geographic Map Analysis and Visualization,” in International Conference on Advances in Information and Communication Technology, Springer, 2024, pp. 127-136, doi: 10.1007/978-3-031-80943-9_14.
Goong, “Top 4 map APIs in the Vietnamese market,” (in Vietnamese), 2023. [Online]. Available: https://goong.io/top-4-loai-map-api-tot-nhat-tai-thi-truong-viet-nam/. [Accessed Feb. 05, 2026].
G. Boeing, “OSMnx: A Python package to work with graph-theoretic OpenStreetMap street networks,” Journal of Open Source Software, vol. 2, no. 12, 2017, doi: 10.21105/joss.00215.
OpenStreetMap contributors, “Copyright and License,” OpenStreetMap, 2017. [Online]. Available: https://www.openstreetmap.org/copyright. [Accessed Feb. 05, 2026].
OpenStreetMap Foundation, “Attribution Guidelines,” OpenStreetMap Foundation Wiki, 2021. [Online]. Available: https://osmfoundation.org/wiki/Licence/Attribution_Guidelines. [Accessed Feb. 05, 2026].
OpenStreetMap Wiki contributors, “Overpass API,” OpenStreetMap Wiki, 2024. [Online]. Available: https://wiki.openstreetmap.org/wiki/Overpass_API. [Accessed Feb. 05, 2026].
