The Forschungszentrum Informatik is developing an emulation system which takes into account pro- blems arising from the concurrent planning of multi agents. Multi agents cause rapid changes in their common world, and the goals of one agent often conflict with those of another. Therefore, we use opportu- nistic planning to model the multi agent task. In this model, events trigger a distributed and con- current planning algorithm. The basic concept consists of a set of behaviours composed of primi- tives which lead to higher-order goals. Autonomo- us multi agents are coordinated and synchronized amongst each other and with their related objects by the communication process and logical sen- sors. Resources evaluate all possible behaviours and monitor the progress of the selected behaviour. The exchange of relevant information depends on the situation and is guided by a priori defined schemata (roles). The incorporation of the individual into a group (e.g. convoy) only succeeds, if the agent adapts its individual behaviours to fit the goals of the group. This loss of freedom for the individual is compensated by its increased safety in the group. Our application arises from traffic engineering. In this context, our demonstrator project is a fra- mework for testing and visualizing the behaviour of multi agents, used to evaluate implemented be- haviour schemes.