Learning Agent Models in SeSAm

Designing the agent model in a multiagent simulation is a challenging task due to the generative nature of such systems. In this contribution we present an extension to the multiagent simulation platform SeSAm, introducing a learningbased design strategy for building agent behavior models.