An Architecture for Integrating LLM-Based Agentic Systems with Multi-Agent Simulations

• Bruno Cascaes Alves UFPel
• Míriam Blank Born UFPel
• Marilton Sanchotene de Aguiar UFPel

DOI: https://doi.org/10.5753/eniac.2026.30634

Resumo

This paper presents an architecture for integrating LLM-based agentic systems with multi-agent simulation platforms, addressing the limitations of traditional, strictly rule-based models. The system couples spatial simulation with context-aware decision-making through a modular, decoupled design. A dedicated API asynchronously manages communication, orchestration, and memory persistence among the simulation environment, a relational database, and the agentic framework. The system’s feasibility is demonstrated in a reference implementation that combines the GAMA simulator with the Agno framework. The analysis of the results reveals that the approach significantly enhances the agents’ adaptability to dynamic scenarios, although it imposes a considerably higher computational cost due to the inference processes required by the LLMs. Consequently, this study highlights a clear trade-off between computational performance and behavioral adaptability while providing a flexible and interoperable infrastructure.

Referências

Agno Team (2024). Agno. Available at: [link]. Accessed: 2026-05-26.

Alves, B., Caldas, L., Junior, U. F., Santos, A., Grossmann, R., Born, M., and Aguiar, M. (2025). Sistema multiagente para simulação de propagação viral no contexto da covid-19 utilizando a ferramenta netlogo. In Anais do XIX Workshop-Escola de Sistemas de Agentes, seus Ambientes e Aplicações, pages 324–331, Porto Alegre, RS, Brasil. SBC.

Alves, B. C., Born, M. B., and Aguiar, M. S. d. (2026). Integrating large language models as cognitive agents into the gama platform for urban mobility simulation. The International FLAIRS Conference Proceedings, 39(1).

Bayer, M. (2024). Sqlalchemy. [link]. Accessed: 2026-05-26.

Dorri, A., Kanhere, S. S., and Jurdak, R. (2018). Multi-agent systems: A survey. IEEE Access.

Olszewski, R., Pałka, P., Turek, A., Kietlińska, B., Płatkowski, T., and Borkowski, M. (2019). Spatiotemporal modeling of the smart city residents’ activity with multi-agent systems. Applied Sciences, 9(10).

Ozdemir, S. (2023). Quick Start Guide to Large Language Models: Strategies and Best Practices for Using ChatGPT and Other LLMs. Addison-Wesley, Boston, MA, USA.

Ramírez, S. (2018). FastAPI. Available at: [link]. Accessed: 2026-05-26.

Richardson, L. and Ruby, S. (2007). RESTful Web Services. O’Reilly Media, Sebastopol, CA, 1 edition.

Sapkota, R., Roumeliotis, K. I., and Karkee, M. (2025). Ai agents vs. agentic ai: A conceptual taxonomy, applications and challenges. Information Fusion, 126.

Stonebraker, M. and Rowe, L. A. (1990). The postgres next-generation database management system. Communications of the ACM, 33(10):78–92.

Taillandier, P., Gaudou, B., Grignard, A., Huynh, Q.-N., Marilleau, N., Caillou, P., Philippon, D., and Drogoul, A. (2019). Building, composing and experimenting complex spatial models with the GAMA platform. Geoinformatica, 23(2):299–322.

Venkadesh, P., Divya, S. V., and Kumar, K. S. (2024). Unlocking AI Creativity: A multi-agent approach with CrewAI. Journal of Trends in Computer Science and Smart Technology, 6(4):338–356.

Wilensky, U. (1999). NetLogo. Center for Connected Learning and Computer-Based Modeling, Northwestern University.

Wu, Q., Bansal, G., Zhang, J., Wu, Y., Li, B., Zhu, E., Li, C., Zhang, L., Rowe, M., Jiang, Z., et al. (2023). AutoGen: Enabling next-gen LLM applications via multi-agent conversation framework. arXiv preprint arXiv:2308.08155.