Applications of large language models for imitating human behavior in an airport
Аuthors
Moscow Aviation Institute (National Research University), 4, Volokolamskoe shosse, Moscow, А-80, GSP-3, 125993, Russia
e-mail: d.g.mchedlishvili@yandex.ru
Abstract
The proposed work is dedicated to the issues of large language models applications in agent modelling. Classic models and approaches allow to create systems, capable of imitating human behavior with high precision, however some difficulties manifest if agent is requested to perform unconventional actions. Applying large language models to this problem allows to circumvent this issue and create decision making system capable of reacting to its environment and current status without having to prepare for every single possible outcome. This approach allows the system to become much more flexible, than its previous counterparts, even if it isn’t capable to completely solve the problem of unreliable reproducibility. For this purpose, an abstract environment of an airport had been created, where LLM agents are tasked to pass every procedure starting from entering the airport all the way until boarding the plane and vice versa. Such environment, being similar enough to an actual airport, is also simple enough for turning complex tasks into a sequence of strings, describing places to which agents have to proceed. Those strings are connected into a single long sentence and then are compared to actual human behavior, collected via a series of polls and written in a similar fashion using Levenstein distance. As a result of the experiments, the agents successfully completed every required task, only relying on basic information about airport and its current state, however they also showed several issues, such as attempting to re-visit places agents have already visited and occasional omitting of non-crucial tasks.
Keywords:
mathematical modelling, large language models (LLM), agent modelling, airportsReferences
- Joon Sung Park, Joseph C. O'Brien, Carrie J. Cai, Meredith Ringel Morris, Percy Liang, Michael S. Bernstein, 2023, Generative Agents: Interactive Simulacra of Human Behavior, https://doi.org/10.48550/arXiv.2304.03442
- Se-eun Yoon, Zhankui He, Jessica Echterhoff, and Julian McAuley. 2024. Evaluating Large Language Models as Generative User Simulators for Conversational Recommendation. In _Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 1490–1504, Mexico City, Mexico. Association for Computational Linguistics. doi:10.18653/v1/2024.naacl-long.83
- Thirunavukarasu, A.J., Ting, D.S.J., Elangovan, K. _et al._ Large language models in medicine. _Nat Med_ 29, 1930–1940 (2023). https://doi.org/10.1038/s41591-023-02448-8
- Taicheng Guo & Xiuying Chen & Yaqi Wang & Ruidi Chang & Shichao Pei & Nitesh V. Chawla & Olaf Wiest & Xiangliang Zhang, 2024, Large Language Model based Multi-Agents: A Survey of Progress and Challenges, https://doi.org/10.48550/arXiv.2402.01680
- Gao, C., Lan, X., Li, N., 2024, Large language models empowered agent-based modeling and simulation: a survey and perspectives. Humanit Soc Sci Commun 11, 1259. https://doi.org/10.1057/s41599-024-03611-3
- Pablo Hernandez-Leal, Bilal Kartal, Matthew E. Taylor A survey and critique of multiagent deep reinforcement learning, DOI: 10.1007/s10458-019-09421-1
- Sanders, Grégory & Mohammadi Ziabari, Seyed Sahand & Mekic, Adin & Sharpanskykh, Alexei., 2021, Agent-Based Modelling and Simulation of Airport Terminal Operations Under COVID-19-Related Restrictions, doi:10.1007/978-3-030-85739-4_18.
- Collins, Andrew, Koehler, Matthew and Lynch, Christopher, 2024, 'Methods That Support the Validation of Agent-Based Models: An Overview and Discussion' Journal of Artificial Societies and Social Simulation 27 (1) 11, doi: 10.18564/jasss.5258
- Gürcan Ö., 2024, LLM-Augmented Agent-Based Modelling for Social Simulations: Challenges and Opportunities , doi: 10.48550/arXiv.2405.06700
- Ayush Chopra & Shashank Kumar & Nurullah Giray-Kuru & Ramesh Raskar & Arnau Quera-Bofarull, 2024, On the limits of agency in agent-based models, https://doi.org/10.48550/arXiv.2409.10568
Download