Langer, P., Nowak, M. A. & Hauert, Ch. (2008) "Spatial Invasion of Cooperation"

Immediate mention of "The evolutionary puzzle of cooperation" started me off nervous, but they end up citing DS Wilson and so apparently have healthy view of cooperation. But they don't sound like it:"Under Darwinian slelection, the evolution of cooperation is a conundrum, whereas non-cooperation (or defection) is not. In the absense of supppporting mechanisms, cooperators perform poortly and decreasein abundance. "

Everything above is true, but it reflects a view heavily influenced by the simplifying assumptions of the models. For example "supporting mechanisms" imply unnaturalness when the truth is that interaction is the rule, not the exception.

But their work seems to acknoweldge this. Taking the same simulations andputting them on a lattice introduces enough inefficiency into communication and well mixed interaction to allow, under many circumstances, selection for altruism. Does 'spatially structured population' mean 'existing in space'?

They do cite a few examples where spatial structure is detrimental to cooperation (Hauert and Doebeli).

One really neat result is that they got spots, or clusters of somewhat uniform size across the simulation depending on the parameters. The explanation for this is that defectors straddling populations can benefit from both and survive.All the action is onthe borders.

Read Wilson and Sober 1994 Reintroduing group selection to the human behavioral sciences.

They emphasized the minimal 3x3 square as the basic unit of cooperation but didn't show or attempt to argue that this is, in some sense, a finite size effect. If they had fine-grained the 3x3 into 6x6 and shown growth they would have demonstrated this effect of simulation-resolution.

Very recent paper. Can't tell if that is good orbad. It means I'm not too late. Econ could use this stuff. Makes me glas I'm not in game theory.

[{Category: 2008]] [{Category: Papers to read]]