Introducing intelligent agents into massively multi-user persistent worlds

In this work we propose an architecture and methodology to add virtual agents to Massively Multi-User Persistent Worlds. It is based on a restricted class of Multi-Agent Systems that meets the requirements of the problem and simplify the implementation process. To demonstrate the application of our methodology we add a set of virtual agents to a 3D chat. The agents survey the users to obtain data to evaluate the processes and components of the virtual world.