Agents are charactized by properties like autonomy, concurrency, flexibilty, intelligence, and mobility. Multiagent systems are operating in open, dynamic environments. Relations between agents, modeling, and simulation are manifold:

1. Multiagent systems can be used as a metaphor to model the system under study. In this case the system is perceived as a community of interoperating individuals. This perception, which related to individual-based or micro-modelling, finds increasingly attention in areas like economy, demography, ecology, sociology, and biology.
   
2. Often software systems are designed as multiagent systems. As simulation systems are software systems, multi-agents as design metaphor are also exploited for designing distributed, parallel simulation systems and  systems for online simulation.
   
3. If a sofware shall be designed as multi-agent system, it is supposed to function in an open and dynamic environment. The software needs to be tested. For this simulation systems are used. The simulated environment generates dynamically (in interaction with the software under test) test cases for that purpose and allows to assess the performance of the designed strategies.