Personalities, physiology, institutions and genetics : simulating ancient societies with intelligent virtual agents

3D Virtual Reality simulations of ancient societies represent a powerful mechanism for combining the knowledge of many researchers (e.g. archaeologists, historians and anthropologists) in a way that it becomes accessible to general audiences, as well as suitable for cross-disciplinary academic collaboration. Through such simulations people with little knowledge in the aforementioned academic disciplines can experience daily lives of people and societies that no longer exist and better understand their cultures. Simulating human activity in all its diversity is a costly and time-consuming exercise comparable in cost and efforts to producing a commercial video game, involving years of development and millions of dollars in funding. Here we propose an approach that can help to significantly decrease the cost and time required for building such simulations, while still capturing enough detail to be useful. By utilising genetic algorithms in combination with computational models of physiological motivation and personalities and constraining virtual agent behaviour through deployment of a formal specification of social norms we can automate the process of building large virtual soci00eties. The manual portion of creating such societies can be reduced to designing an initial small sample of avatars, specifying social norms of the given society and annotating the virtual environment in a particular way. Most other development stages can be automated. The usefulness of our approach is demonstrated by applying it to simulating everyday life in the ancient city of Uruk, 3000 B.C. and the simulation of daily life in an aboriginal tribe in Australia just before the arrival of the first European settlers.