Abstract for: Combining Agent-Based Modelling and System Dynamics: Hybrid modelling in the scope of artificial intelligence

System Dynamics (SD) and Agent-based Modelling (ABM) are generally seen as two contrasting modeling paradigms. If SD and ABM are used as solitary modelling approaches, this view is inevitability the result. In this paper, hybrid modelling is envisaged by combining both modeling techniques. Although there are some papers that aim to use SD and ABM for hybrid modeling, the research field is quite undiscovered. The paper introduces a generic concept on how to combine System Dynamics and Agent-Based Modelling. One the one hand, it is possible to implement a SD model inside an agent, where the SD structure controls the behavior of the agent. On the other hand, agents can be embedded into a SD model via four different envisaged ways. For the former case, a case study is documented in this paper, where SD is used to model the behavior of virtual humans. In particular – different from the common – SD is used from the bottom-up inside each instance of an agent population living in a virtual environment. From the perspective of artificial intelligence research, it is shown how System Dynamics can be used to design action selection models to provide autonomy to virtual humans.