Human cognition of space involves its structuring and enrichment with complex meaning, in particular related to its action potentials, and thus results in the formation of separate places. Despite its central role, however, work on the representation of place in Geographical Information Systems (GIS) is still at an early stage. Particular challenges arise from place-inherent characteristics, and include vagueness, subjectivity and dynamic variability in terms of location, footprint and semantics. In this project, the use of agent-based models (ABM) is explored to localize places in space. For this, it is necessary to simulate the process of place formation of each individual agent. With a strong focus on its action-related dimension, place is conceptualized as a subset of atomic spatial entities, which are bound by a functional unity condition with regards to a complex action. Based on an hierarchical action model, our framework allows agents to evaluate the suitability of individual geo-atoms with regards to these actions, and choose their behavior accordingly. The usefulness of the approach is demonstrated by simulating pedestrian movement in the city centre of Augsburg, Germany.