A methodology for the development of autonomous agents for virtual environments

Abstract

In the last years, developers have focused on Massively Multi-User Persistent Worlds, This virtual communities host thousands of users interacting with each other through the Internet in "real" time. Users are not the only living objects inside an MMPW. Non-player characters also co-exist with regular users, and need their own social behavior. This behavior must be described at a semantic level, a level that allows control of the environment's complexity and dynamic evolution. In this work, we propose an architecture and methodology to introduce autonomous agents into Virtual Environments. It is based on a new class of Multi-Agent Systems, which is named Moderately Open Multi-Agent Systems. The HABA.DM methodology for the analysis and design of systems, the HABA.PL programming language, and the HABA.PM project manager are the structural elements of the HABA development process we propose. Moreover, we describe and demonstrate the use of our model and development framework to build other typical systems of Internet computing; for example, monitoring and control systems and e-commerce systems.