Visualizing agent-based simulation dynamics in a CAVE — Issues and architectures

Displaying an agent-based simulation on an immersive virtual environment called CAVE (Cave Automatic Virtual Environment), a human expert is enabled to evaluate the simulation's dynamics from the same point of view as in real life — from a within perspective instead of a birds eye view. As this form of face validation is useful for many multiagent simulations, it should be possible to setup such a system with as little effort as possible. In this context, we systematically analyse the critical issues that a realization of such a system raises. Addressing these problems, we finally discuss design aspects of basic framework architectures.