Elamur Research⌕
☰
home / papers / 2026 / MobileCity: An Efficient Framework for Large-Scale Urban Behavior Simulation

MobileCity: An Efficient Framework for Large-Scale Urban Behavior Simulation

Paper · 2026 · Xiaotong Ye, Nicolas Bougie, Toshihiko Yamasaki et al.

Summary

Generative agents powered by LLMs can imitate everyday human conduct. However, current urban simulators give agents fixed personas, ignore routines, duties and weather, assume a single simplistic way of travelling, and call the model synchronously with lengthy dialogues. This makes populations of thousands too expensive to run. MobileCity, from Woven by Toyota and the University of Tokyo, addresses these gaps with a tile-based virtual town whose agents are seeded from questionnaire-derived demographic and personality profiles. Each agent's choices combine three drivers: needs that decay over time, habits tied to familiar times of day that gradually fade, and calendar obligations that exclude actions clashing with upcoming mandatory tasks. Agents also choose how to travel, factoring in weather, temperature and venue opening hours. Efficiency comes from four design choices: the LLM picks options instead of writing free text, agents swap memory indices rather than conversing, inference is batched and asynchronous, and only events are logged. In a 4,000-agent run, the authors report more realistic behaviour than baselines at low computational cost. This makes the system a practical tool for planning and mobility forecasting.

Key points

  • Each agent tracks eight needs mapped to Maslow's hierarchy. Decay rates differ by persona and need type, and an importance vector weights each need. Habits follow a circular Gaussian over the daily cycle with Ebbinghaus-style forgetting.
  • Agents can get around in one of three ways: on foot, by bus, or on a PMV (personal mobility vehicle).
  • The scalability techniques are:
    • multiple-choice action selection, where the model returns only an option index;
    • memory-index exchange with relationship-score updates in place of generated dialogues;
    • asynchronous batched inference using per-agent local clocks;
    • event-level logging in OpenSearch.
  • The simulated share of travel time closely matches a proprietary ground-truth dataset across demographic groups. Walking accounts for roughly 90–97%, bus for about 2–10%, and PMV stays near zero, being especially rare in rainy conditions.
  • The paper appeared in the EACL 2026 Industry Track, and the code is released at github.com/Tony-Yip/MobileCity.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓The University of TokyoTokyo · Japan90 works · 9,156 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Social Sciences · 1,772 worksEngineering · 133 worksBuilding and Construction · 5 worksTransportation · 6 worksOcean Engineering · 1 work

Cite

Cite as Xiaotong Ye, Nicolas Bougie, Toshihiko Yamasaki and Narimawa Watanabe. 2026. MobileCity: An Efficient Framework for Large-Scale Urban Behavior Simulation. Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.eacl-industry.21

@inproceedings{ye2026mobilecity,
  title = {{MobileCity: An Efficient Framework for Large-Scale Urban Behavior Simulation}},
  author = {Xiaotong Ye, Nicolas Bougie, Toshihiko Yamasaki and Narimawa Watanabe},
  booktitle = {Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track)},
  year = {2026},
  doi = {10.18653/v1/2026.eacl-industry.21},
  url = {https://doi.org/10.18653/v1/2026.eacl-industry.21}
}

References

31 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Murray R Barrick and Michael K Mount. 1991. The big five personality dimensions and job performance: a meta-analysis. Personnel psychology, 44(1):1–26. find
  2. Nicolas Bougie and Narimasa Watanabe. 2024. Generative adversarial reviews: When llms become the critic. arXiv preprint arXiv:2412.10415. — [arXiv:2412.10415](https://arxiv.org/abs/2412.10415) find
  3. Nicolas Bougie and Narimasa Watanabe. 2025. CitySim: Modeling urban behaviors and city dynamics with large-scale LLM-driven agent simulation. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 215– 229, Suzhou (China). Association for Computational Lingu… find
  4. Lars Böcker, Martin Dijst, and Jan Prillwitz. 2013. Impact of everyday weather on individual daily travel behaviours in perspective: A literature review. Transport Reviews, 33. find
  5. Walter B. Cannon. 1932. The Wisdom of the Body. W. W. Norton & Company. find
  6. Weize Chen, Yusheng Su, Jingwei Zuo, Cheng Yang, Chenfei Yuan, Chi-Min Chan, Heyang Yu, Yaxi Lu, Yi-Hsin Hung, Chen Qian, Yujia Qin, Xin Cong, Ruobing Xie, Zhiyuan Liu, Maosong Sun, and Jie Zhou. 2024. Agentverse: Facilitating multi-agent collaboration and exploring emergent behaviors. In Internatio…
  7. Jie Feng, Jun Zhang, Junbo Yan, Xin Zhang, Tianjian Ouyang, Tianhui Liu, Yuwei Du, Siqi Guo, and Yong Li. 2024. Citybench: Evaluating the capabilities of large language model as world model. arXiv preprint arXiv:2406.13945. — [arXiv:2406.13945](https://arxiv.org/abs/2406.13945) find
  8. Leon Festinger. 1957. A Theory of Cognitive Dissonance. Stanford University Press. find
  9. Zhao Kaiya, Michelangelo Naim, Jovana Kondic, Manuel Cortes, Jiaxin Ge, Shuying Luo, Guangyu Robert Yang, and Andrew Ahn. 2023. Lyfe agents: Generative agents for low-cost real-time social interactions. CoRR, abs/2310.02172. — [arXiv:2310.02172](https://arxiv.org/abs/2310.02172) find
  10. Munyeong Kim and Sungsu Kim. 2023. Generative AI in mafia-like game simulation. CoRR, abs/2309.11672. — [arXiv:2309.11672](https://arxiv.org/abs/2309.11672) find
  11. Michael R King and ChatGPT. 2023. A conversation on artificial intelligence, chatbots, and plagiarism in higher education. Cellular and molecular bioengineering, 16(1):1–2. find
  12. Junkai Li, Siyu Wang, Meng Zhang, Weitao Li, Yunghwei Lai, Xinhui Kang, Weizhi Ma, and Yang Liu. 2024. Agent hospital: A simulacrum of hospital with evolvable medical agents. CoRR, abs/2405.02957. — [arXiv:2405.02957](https://arxiv.org/abs/2405.02957) find
  13. Aman Madaan, Shuyan Zhou, Uri Alon, Yiming Yang, and Graham Neubig. 2022. Language models of code are few-shot commonsense learners. In Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1384–1403. find
  14. Abraham Harold Maslow. 1943. A theory of human motivation. Psychological review. find
  15. OpenSearch Project. 2021. Opensearch. https://github.com/opensearch-project/OpenSearch. GitHub repository. ↗
  16. Joon Sung Park, Joseph C. O’Brien, Carrie Jun Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein. 2023. Generative agents: Interactive simulacra of human behavior. In The 36th Annual Symposium on User Interface Software and Technology (UIST), pages 2:1–2:22. find
  17. Joon Sung Park, Carolyn Q. Zou, Aaron Shaw, Benjamin Mako Hill, Carrie J. Cai, Meredith Ringel Morris, Robb Willer, Percy Liang, and Michael S. Bernstein. 2024. Generative agent simulations of 1,000 people. CoRR, abs/2411.10109. find
  18. Jonathan Gershuny. 2003. Changing times: Work and leisure in postindustrial society. OUP Oxford. find
  19. Fritz Heider. 1958. The Psychology of Interpersonal Relations. Wiley. find
  20. Sirui Hong, Mingchen Zhuge, Jonathan Chen, Xiawu Zheng, Yuheng Cheng, Jinlin Wang, Ceyao Zhang, Zili Wang, Steven Ka Shing Yau, Zijuan Lin, Liyang Zhou, Chenyu Ran, Lingfeng Xiao, Chenglin Wu, and Jürgen Schmidhuber. 2024. Metagpt: Meta programming for A multi-agent collaborative framework. In Inter…
  21. Amanda L Rebar, Ryan E Rhodes, and Benjamin Gardner. 2019. How we are misinterpreting physical activity intention–behavior relations and what to do about it. International Journal of Behavioral Nutrition and Physical Activity. find
  22. David C Rubin and Amy E Wenzel. 1996. One hundred years of forgetting: A quantitative description of retention. Psychological review. find
  23. Vinay Samuel, Henry Peng Zou, Yue Zhou, Shreyas Chaudhari, Ashwin Kalyan, Tanmay Rajpurohit, Ameet Deshpande, Karthik Narasimhan, and Vishvak Murahari. 2024. Personagym: Evaluating persona agents and llms. arXiv preprint arXiv:2407.18416. find
  24. Daniel Tischner. 2018. Multi-modal route planning in road and transit networks. CoRR, abs/1809.05481. — [arXiv:1809.05481](https://arxiv.org/abs/1809.05481) find
  25. Zhilin Wang, Yu-Ying Chiu, and Yu Cheung Chiu. 2023. Humanoid agents: Platform for simulating human-like generative agents. In Conference on Empirical Methods in Natural Language Processing (EMNLP). find
  26. Mark Wardman and Jeremy Toner. 2020. Is generalised cost justified in travel demand analysis? Transportation, pages 75–108. find
  27. Wendy Wood, Asaf Mazar, and David T Neal. 2022. Habits and goals in human behavior: Separate but interacting systems. Perspectives on Psychological Science. find
  28. Yuwei Yan, Qingbin Zeng, Zhiheng Zheng, Jingzhe Yuan, Jie Feng, Jun Zhang, Fengli Xu, and Yong Li. 2024. Opencity: A scalable platform to simulate urban activities with massive llm agents. arXiv preprint arXiv:2410.21286. — [arXiv:2410.21286](https://arxiv.org/abs/2410.21286) find
  29. Yangbin Yu, Qin Zhang, Junyou Li, Qiang Fu, and Deheng Ye. 2024. Affordable generative agents. CoRR, abs/2402.02053. — [arXiv:2402.02053](https://arxiv.org/abs/2402.02053) find
  30. An Zhang, Yuxin Chen, Leheng Sheng, Xiang Wang, and Tat-Seng Chua. 2024. On generative agents in recommendation. In Conference on Research and Development in Information Retrieval (SIGIR), pages 1807–1817. find
  31. Wangchunshu Zhou, Yuchen Eleanor Jiang, Long Li, Jialong Wu, Tiannan Wang, Shi Qiu, Jintian Zhang, Jing Chen, Ruipu Wu, Shuai Wang, Shiding Zhu, Jiyu Chen, Wentao Zhang, Ningyu Zhang, Huajun Chen, Peng Cui, and Mrinmaya Sachan. 2023. Agents: An open-source framework for autonomous language agents. C… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ye-2026-mobilecity-efficient-framework-large-scale/