Simulating human behaviors in agent societies

As increasing numbers of processors and agents pervade the human environment, societies comprising both humans and agents will emerge. Presently, it is unknown how a person might fair in such mixed societies. For the societies to operate effectively and efficiently, it is important for the humans and agents to recognize and understand each other’s behavior. This paper provides an initial step in that understanding via two contributions: (1) we provide models, within a limited domain, for agents that behave like humans and (2) we present the results of simulated interactions between the human-like agents and a variety of purely rational agents. Our models for the behaviors of people are based on recent sociological research by Simpson and Willer [10] that explores humans’ cooperative prosocial behavior, a conceivably non-rational process. Modeling human behaviors presents a means of exploring and understanding motivations, consequences, and resolutions to human-agent interactions. We aspire to exploit measured human behavior in order to observe its ramifications in an agent world, and to motivate development of human-agent societies. Our results show that, although there are pitfalls to which humans are vulnerable, there exist niches for human prosperity in a rational agent world.