The Evolution of Cooperation

From Wikipedia, the free encyclopedia

The Evolution of Cooperation is a 1984 book and a 1981 article of the same title by political science professor Robert Axelrod. The nine-page article is currently one of the most cited articles ever to be published in the journal Science ([1]).

In it, Axelrod explores the conditions under which fundamentally selfish agents will spontaneously cooperate. To perform this study, Axelrod developed a variation of prisoner's dilemma (PD), involving repeated PD interactions between two players (i.e., strategies written as computer programs) in a computerised tournament. This iterated prisoner's dilemma (IPD) format, he found, tends to offer a long-term incentive for cooperation, even though there is a short-term incentive for defection (the opposite of cooperation).

Axelrod invited academic colleagues all over the world to devise strategies to compete in an IPD tournament. The results ranged in many variables: algorithmic complexity, initial hostility, capacity for forgiveness, etc. After an initial tournament that simply compared pairs of strategies for success when paired in an IPD, Axelrod arranged a meta-tournament where strategies represented sub-populations in a large population of agents, and an agent could switch to another strategy if it noticed that one of its neighbors was using that strategy with greater success than its own. It should also be noted the simplest system, Tit for Tat, won the tournament.

Tit-for-tat had a number of important features as a strategy - it was "NICE" (it didn't defect first), and it was "provocable" (it fought back if it was attacked). Tit-for-tat never did better than its immediate opponent, but was able to cooperate very well with itself and with other "NICE" strategies - thereby harvesting the substantial benefits of mutual cooperation. Ironically, more fierce strategies tended to "canniblize" each other leading to fewer gains. They also could not take excessive advantage of Tit-for-Tat other than in their initial suprize defection - because Tit-for-Tat retaliated. (Tit-for-tat has been described as the "silver rule"). When Tit-for-tat represented a large enough proportion of the population, other "NICE" strategies could also effectively co-habitate.

The book included two chapters comparing Axelrod's findings to surprising findings in seemingly unrelated fields. In one of these, Axelrod examined spontaneous instances of cooperation during trench warfare in World War I. Troops of one side would shell the other side with mortars, but would often do so on a rigid schedule, and aim for a specific point in the other side's trenches, allowing the other side to minimize casualties. The other side would reciprocate in kind. The generals on both sides were satisfied that shelling was occurring and therefore the war was progressing satisfactorily, while the men in the trenches found a way to cooperatively protect themselves.

Considerable additional work has been done in this area. Repeated/Iterated prisoner's dilemma is one of the most important areas of game theory with implications for all the social sciences and for practical government policy. One of the key research findings in this area is the FOLK THEOREM - please see :

     http://en.wikipedia.org/wiki/Talk:Folk_theorem_%28game_theory%29

     http://en.wikipedia.org/wiki/Folk_theorem_%28game_theory%29

[edit] References

• Axelrod, Robert. (1984). The Evolution of Cooperation. New York: Basic Books, ISBN 0-465-02121-2
• Axelrod, Robert. (2006). The Evolution of Cooperation Revised edition Perseus Books Group, ISBN 0-465-00564-0
• Axelrod, Robert. (1981). "The Evolution of Cooperation." Science, 211(4489):1390-6