Agent Based Modelling and Simulation - applicability within OR at FFI

Within the Military Operations Research community, simulation models constitute a crucial part of the analysis tool box. These models are built in order to enhance the understanding of a specific system. The system can for instance be attrition based land warfare or technical availability of aircrafts. So far these models have been concerned mainly on the technological part of the problems. Nowadays, low intensity operations constitute an increasing amount of the relevant military operations. Within these modern operations human decision-making and human behaviour comprise a much more important part compared to the cold-war scenarios. The objectives of a military operation nowadays are often about reducing the tension in a conflict, while the earlier objectives were concerned on attrition. This means that the relative importance between technology and human behaviour has changed a lot, in order to determine success or not in a military operation. Agent based models represent an alternative way to describe military operations. Still we talk about simulation models, but the agent based models typically put the human decision making in focus at the expense of high fidelity within technology representations. Another difference is the input level: Agent based models apply the input to describe each agent (i.e. low level), while the higher level behaviour constitutes the output from the model. On the other hand, traditional simulation models typically involve scripted behaviour which means that the top level behaviour is given by the input parameters. So far FFI has not applied agent based models within Operations Research. Nevertheless, a lot of similar organizations do, and this report describes a variety of their applications. The trend found within these applications tells that the most promising ongoing activities are focused on human centric applications, i.e. applications were human behaviour and human decision making are deliberately represented.