On the Use of Spatial Games in Explaining Human Cooperation
Evolutionary Computation (CEC), 2016 IEEE Congress on
Spatial games are extensively used to study how cooperation evolves in human populations. Nevertheless, spatial games have several limitations which can produce misleading results. Specifically, the regular lattice structure creates artificial interactions and the reliance on a Moran process updating, coupled with weak selection, makes it difficult to switch strategies. These problems contribute to over-estimation of cooperation levels when compared to human experiments. In this paper these issues are discussed in depth. Two theorems relating to Moran process updating in spatial games are included.
Locate the Document
G. W. Greenwood, "On the use of spatial games in explaining human cooperation," 2016 IEEE Congress on Evolutionary Computation (CEC), Vancouver, BC, 2016, pp. 351-356.