Evolutionary Game Theory Reevaluated

According to the basic models of evolutionary game theory that we studied, selfish strategies are more likely to prevail over cooperative strategies, even if they are not social optimal.  Biologists have previously believed that cooperative behaviors prevail only when they are also the best response for the individual, but there are new theories to suggest that animals engage in socially optimal behavior even if it is not in their immediate best interest.  It seems that some animals understand that it is in their long term best interest (and in the interest of their group) to engage in behaviors such as warning calls and sharing food, even if it leaves the giver slightly worse off.

Though a prisoner’s dilemma is traditionally a one-time game, cooperative decisions have to be made over and over.  Choosing to not make a warning call may be the best decision in an individual moment, but over time the optimal strategy changes to one of cooperation.  This type of repeated game where one player mimics the previous move of the other player is called ‘tit-for-tat’.

In the real animal world, these strategies look a little more complicated; fitness has to be measured on a gradation, and animals do occasionally forgive the betrayal of those in their group.  In this case, these researchers realized that one player can defect a certain percentage of the time and it is still in the other’s best interest to cooperate.  In these repeated games, the player who is a periodic extortionist has a strictly higher payoff than the cooperative player.  In the long run though, the generous strategy prevails because other members employ the ‘tit for tat’ idea and also generally play generous strategies, making each individual and the group as a whole better off.

http://www.scientificamerican.com/article/game-theory-calls-cooperation-into-question1/

Advertisements

1 thought on “Evolutionary Game Theory Reevaluated”

  1. I find it interesting that tit-for-tat offers players the ability to defect from the mutually beneficial strategy only to return to it without a problem. I feel like in a real world application after so many times one player or another would get frustrated and the repeated game would end.

Comments are closed.