The Construction of Agent Simulations of Human Behavior
A general definition of the minimum required structure of computer agents that are capable of accurately representing human behavior in agent-based models is offered. Included is the abstract definition of the Environment, the State Vector, the Perceptor, the Actor, and the Ratiocinator components of the agent structure and their interaction. The synthetic population and associated incident distributions are then defined. Finally, practical considerations of time, accuracy and path dependency are examined.
KeywordsHuman agent model Synthetic population Narrative Ratiocinator Memory state vector Virtual markets Path dependency Structural limits
- 1.Parker, R.A., Bakken, D.: Predicting the unpredictable: agent-based models in market research. In: Mouncey, P., Wimmer, F. (eds.) Best Practices in Market Research. Wiley, Hoboken (2007)Google Scholar
- 2.Parker, R.A., Perroud, D.: Exploring markets with agent-based computer modeling. In: Proceedings of the 61st ESOMAR International Congress, Montreal, Canada (2008)Google Scholar
- 5.Conway, J.: Mathematical games: the game of life. In: Gardner, M. (ed.) Scientific American (1970)Google Scholar
- 7.Parker, R.A.: A demonstration of AirVM. In: Proceedings of the AGIFORS Schedule and Strategic Planning Study Group, Lausanne, Switzerland (2010)Google Scholar
- 11.Parker, R.A., Holmes, B.: Using predictive analytics for corporate shuttle decisions. In: Proceedings of the NBAA Convention, Las Vegas, Nevada, USA (2015)Google Scholar