Emergent Behavior as a Function of Information

In this paper, I briefly introduce and summarize the state of the art of emergent behavior. I also present my own experiment and findings. I used a competitive domain to investigate how information contributes to emergent behavior. Specifically, I attempt to see if there is a certain required level of information for emergent behavior to arise. By using the competitive domain to decide “which team is stronger”, I simply examine the results of a round-robin tournament between teams of agents. The only difference between each team is simply how far the member agents can see. My hypothesis was that there would be a certain “cutoff” point, where after a certain sight range teams, having access to enough or more than the sufficient level of information, would act dramatically better. Sadly, this hypothesis was most likely falsified by the evidence. What I did find was that both too much and too little sight range creates weaker teams, while a middling sight range creates the strongest. This paper ends with a discussion of what next steps might be taken in investigating emergent behavior in this fashion.