We present two classes of agent-based complex systems, namely traffic flow dynamics and spatial evolutionary games. We first discuss the individual-based interactions and dynamics mechanisms and their connections with statistical mechanics models. Furthermore, we derive mesoscopic integro-differential equations as approximations of the agent-based models which reveal the relationship between the microscopic individual behaviors and macroscopic morphologies.
Finally, we discuss hierarchical mesoscale models that allow variable granularity descriptions of the underlying complex systems.