One of the most widely used techniques for the analysis of crowds such as evacuations, pedestrian flow or vehicular traffic among others is agent-based simulation. Reasoning and decision making are crucial processes but modeling them correctly requires knowledge representation; the methodologies used to represent knowledge and manage interactions between agents are fundamental to creating effective and realistic models. Multi-agent models often oversimplify human or social behaviors, which can result in unrealistic representations. The lack of modeling accuracy and the complex interaction between agents can limit the validity of the results, being one of the biggest challenges in multi-agent models.

We propose making use of the collective intelligence genome proposed by Thomas Malone as a methodology for the representation of crowds through agent-based simulation. The definitions that compose it can be understood as an ontology for crowd representation.

We have related the components of the genome directly to the components of an agent-based simulation model: What=agent actions, why=behavior motivators, how = how they perform actions and describe it as an ontology. We have also developed case studies to test the methodology.