Oscillations in the evolution of reciprocity

Karl Sigmund
1989-01-01
Abstract:An increasing amount of attention has been paid to the Iterated Prisoner's Dilemma as a game theoretical paradigm for the evolution of co-operation based on reciprocity. In repeated encounters, two individuals are faced with the options to co-operate or to defect (C'or ID). Joint co-operation leads to a payoff R (reward) which is higher than the payoff P (punishment) for mutual defection. But if one player “cheats” by defecting while the other co-operates, then his payoff T (temptation) is larger than R, while the co-operator's payoff S (sucker) is smaller than P. In addition to T> R> P> S one usually assumes R>(S+ T) in order to simplify the analysis by making alternations of co-operation and defection less rewarding than all-out co-operation.If the garne consists of a single encounter (or of a fixed number of encounters known to both players) then the best choice is to defect. But if the length of the game is unknown, as for …
What problem does this paper attempt to address?