Benefits of clustering among the Internet search agents caught in the n-person prisoner's dilemma game

In searching for information on the Internet, often times one experiences contention of information servers. Usually, information on the Internet is collected by autonomous search agents that send out queries to the servers that may have the information sought. From a single agent's perspective, sending out as many queries as possible maximizes the chance of achieving the information sought. However, if every agent does the same, the information sites will be overloaded and most of the agents will be dissatisfied. In general, cooperation-sending a moderate number of queries-is desired for everyone's good. In essence, the Internet search agents are caught in the n-person prisoner's dilemma game. When the number of available information sites is much larger than that of the information-seeking agents (i.e., the resource is abundant), cooperation may not be necessary since there is little incentive to cooperate. However, when the resource is scarce, cooperation will lead more agents successfully retrieving the information within reasonable time. It is, however, generally not possible to know how many information sites are available in the world and how many other agents may seek the same information sites. We present possible benefits of accessing "local" information sites, forming "communities" leading to a global satisfaction of the agents involved.