Injecting Data into Agent-Based Simulation.

Many agent-based models, use -standard distributions in several steps of the design: configuring the initial conditions of simulations, distributing objects spatially, and determining exogenous factors or aspects of the agents΄ behaviour. An alternative approach that is growing in -popularity is data-driven agent-based simulation. This paper encourages -modellers to continue this trend, discussing some guidelines for finding suitable data and feeding models with it. In addition it proposes to merge the principles of microsimulation into the classical logic of agent-based simulation, adapting it to the data-driven approach. A case study comparing the two approaches is provided. [ABSTRACT FROM AUTHOR]