Geo-savvy Agents
Agents in agent-based simulations usually move through space in simplified ways, following predefined paths or shortest routes. Real people find their way by perceiving their surroundings, building mental representations of space, reasoning about where to go, and adapting as they move and interact with others. With the concept of geo-savvy agents, developed together with Franziska Klügl, I bring this knowledge about wayfinding and navigation into the agent-based modelling paradigm. At its core is a cognitive synthesis challenge: how can models of perception, representation, reasoning, action and interaction be connected so that together they produce realistic spatial behaviour? One approach we pursue is concurrent process architectures, in which these cognitive processes run in parallel rather than as a fixed sequence. The work builds on our earlier research on affordances, which models what the environment offers an agent as something that emerges from the interaction between agent and environment. Geo-savvy agents are intended to make simulations of urban systems, mobility and human navigation more realistic and easier to explain.
Geo-savvy agents and concurrent process architectures
- Klügl, F. & Timpf, S. (2026). Geo-Savvy Agents: Integrating Spatial Cognition into Agent-Based Simulation. Poster, August 2026. ResearchGate
- Timpf, S. & Klügl, F. (2026). From Sequential Decisions to Concurrent Processes: Advancing Human Navigation Models for Simulation (Short Paper). In 17th International Conference on Spatial Information Theory (COSIT 2026). LIPIcs, vol. 393, pp. 36:1-36:8, https://doi.org/10.4230/LIPIcs.COSIT.2026.36
- Timpf, S. & Klügl, F. (2026). Realistic Agent Behaviour Based on Concurrent Process Architectures – Modelling Human Navigation. In Multi-Agent Systems, Lecture Notes in Computer Science. Springer, Cham. DOI: 10.1007/978-3-032-22817-8_18
Earlier work on affordances
- Timpf, S. & Klügl, F. (2023). Modelling affordances as emergent phenomena. In 12th International Conference on Geographic Information Science (GIScience 2023), LIPIcs, 72:1–72:6. DOI: 10.4230/LIPIcs.GIScience.2023.72
- Klügl, F. & Timpf, S. (2021). Towards more explicit interaction modelling in agent-based simulation using affordance schemata. In KI 2021: Advances in Artificial Intelligence, 324–337. Springer, Cham. DOI: 10.1007/978-3-030-87626-5_24
- Klügl, F. & Timpf, S. (2018). Approaching interactions in agent-based modelling with an affordance perspective. Lecture Notes in Computer Science 10738, 21–37. DOI: 10.1007/978-3-319-91899-0_2