Articles | Volume 18, issue 15
https://doi.org/10.5194/gmd-18-4983-2025
https://doi.org/10.5194/gmd-18-4983-2025
Model description paper
 | 
15 Aug 2025
Model description paper |  | 15 Aug 2025

InsNet-CRAFTY v1.0: integrating institutional network dynamics powered by large language models with land use change simulation

Yongchao Zeng, Calum Brown, Mohamed Byari, Joanna Raymond, Thomas Schmitt, and Mark Rounsevell

Related authors

Exploring the opportunities and challenges of using large language models to represent institutional agency in land system modelling
Yongchao Zeng, Calum Brown, Joanna Raymond, Mohamed Byari, Ronja Hotz, and Mark Rounsevell
Earth Syst. Dynam., 16, 423–449, https://doi.org/10.5194/esd-16-423-2025,https://doi.org/10.5194/esd-16-423-2025, 2025
Short summary

Cited articles

Acs, S., Hanley, N., Dallimer, M., Gaston, K. J., Robertson, P., Wilson, P., and Armsworth, P. R.: The effect of decoupling on marginal agricultural systems: implications for farm incomes, land use and upland ecology, Land Use Policy, 27, 550–563, https://doi.org/10.1016/j.landusepol.2009.07.009, 2010. 
Ariti, A. T., van Vliet, J., and Verburg, P. H.: The role of institutional actors and their interactions in the land use policy making process in Ethiopia, J. Environ. Manage., 237, 235–246, https://doi.org/10.1016/j.jenvman.2019.02.059, 2019. 
Arnott, R., De Palma, A., and Lindsey, R.: Information and time-of-usage decisions in the bottleneck model with stochastic capacity and demand, Eur. Econ. Rev., 43, 525–548, https://doi.org/10.1016/S0014-2921(98)00013-0, 1999. 
Bainbridge, A., Troppe, T., and Bartley, J.: Responding to research evidence in Parliament: A case study on selective education policy, Rev. Educ., 10, e3335, https://doi.org/10.1002/rev3.3335, 2022. 
Banerjee, S., Agarwal, A., and Singla, S.: Llms will always hallucinate, and we need to live with this, arXiv [preprint], https://doi.org/arXiv:2409.05746, 2024. 
Download
Short summary
Understanding environmental policy interventions is challenging due to complex institutional actor interactions. Large language models (LLMs) offer new solutions by mimicking the actors. We present InsNet-CRAFTY v1.0, a multi-LLM-agent model coupled with a land system model, simulating competing policy priorities. The model shows how LLM agents can simulate decision-making in institutional networks, highlighting both their potential and limitations in advancing land system modelling.
Share