Self-organized norms of behavior under interactions of selfish agents

One of the aspects of life is the desire for self-preservation. In a collective system, which is highly systematized and autonomous, if an individual desires profit and self-preservation, a "social dilemma" develops between the individual goal and that of the system, which requires an output from the agent. In this study, the various "norms of behavior" that lead to an emergent agent society that is able to overcome "social dilemmas" though the interactions between the agents were examined. Agents were not required to recognize their complex environment precisely. Instead, they determined their actions on the basis of an internal evaluation function. The resultant set of such decisions made by the agents was sufficient to drive the social system harmoniously. Self-evaluation of their local interactions by the agents was able to allow both the need for various norms of behavior to coexist and for a harmonious society to be developed. We constructed a competitive social system consisting of selfish autonomous agents. Agents interact with their alternative actions repeatedly as games. Each agent had an independent evaluation table for their actions. The agents' strategies were adapted on the basis of their individual evaluations, and agents' norms of behavior satisfy. In such a system model, the agents divide themselves into groups according to their norms of behavior. Between these groups, there exist relationships such as those that exist in real societies. Furthermore, we suggest a method of controlling autonomous distributed robots with flexible structure using the above mentioned approach.