This week’s PNAS has another (open access!) paper taking a crack at the problem of how cooperation can evolve. The authors create a world where cooperation arises spontaneously in a population of selfish individuals by modeling a fundamental human drive: the desire for a good neighborhood.
Helbing and Yu set up a model world ruled by the Prisoner’s Dilemma, a common game theory scenario in which pairs of interacting individuals can choose to cooperate or not cooperate with each other. If both refuse to cooperate, neither gets anything; if one cooperates and the other doesn’t, the cheater gets a reward, but the cooperator pays a cost; if both cooperate, then they both get a smaller reward. If neither interactor can predict the other’s choice, the most sensible strategy is to just never cooperate – you make out pretty well when the other guy is silly enough to cooperate with you, and you’re no worse off than you started out if you both refuse to cooperate.
Previous models have made cooperation work in Prisoner’s Dilemma situations a few different ways. One way is to allow individuals to remember how they have treated each other over multiple iterations of the PD interaction, so that cheaters can be punished [$-a]; another is to let the game play out across space in such a way that cooperators can cluster together, so that they are more likely to interact with other cooperators [$-a].
Helbing and Yu’s model is a variation on the “spatial” flavor – individuals occupy cells in a grid, and interact with those in adjacent cells. Strictly speaking, it isn’t an evolutionary model (even though the authors describe it as such), because there doesn’t seem to be any inheritance of behavior from one generation to another; instead, individuals “learn” from their neighbors, imitating the ones who are most successful in terms of interaction rewards. There’s a random element to individual behavior, to approximate trial and error strategies. Perhaps most importantly, individuals can migrate across the grid, moving to adjacent unoccupied cells where they expect to find a greater reward.
Neither imitation nor migration alone allow cooperation to survive in this model world, but some interaction between the two does. This result holds, apparently, for a wide range of possible combinations of payoff conditions. For some conditions, the model will even allow cooperators to “invade” a world full of non-cooperators. The speed with which individuals can move across the grid – cooperators seeking other cooperators, and avoiding cheaters – is critical, say the authors. They call this “success-driven migration” – and it does seem to allow cooperation – though not altruism – to arise out of selfishness.
See also Wired Science’s coverage.
M. Doebeli, C. Hauert (2005). Models of cooperation based on the Prisoner’s Dilemma and the Snowdrift game Ecology Letters, 8 (7), 748-66 DOI: 10.1111/j.1461-0248.2005.00773.x
D. Helbing, W. Yu (2009). The outbreak of cooperation among success-driven individuals under noisy conditions PNAS DOI: 10.1073/pnas.0811503106
M.A. Nowak, R.M. May (1992). Evolutionary games and spatial chaos Nature, 359 (6398), 826-829 DOI: 10.1038/359826a0
R.L. Trivers (1971). The evolution of reciprocal altruism Quarterly Rev. Biol., 46, 35-57