Genetic Programming and Common Pool Resource Problems with Uncertainty

When applying an evolutionary computational technique to a problem, Game Theoretic predictions, like the Nash equilibrium, usually describe the problems generated solutions. Human behaviour in the same scenarios has, however, been shown to not conform to these game theoretic predictions. In this paper, a Common Pool Resource (CPR) problem is used to compare the performances of human players and the generated solutions of a Genetic Programming (GP) algorithm. Previously, it has been shown that GP will converge as expected to the Nash Equilibrium predicted behaviour. However, under certain evolutionary scenarios, human-like play appears to emerge. In this paper, the effects of environmental pressures on the behaviours that Genetic Programming generates for a group based CPR dilemma. Similarities are drawn to human behaviours in similar games with a discussion of what influences the evolutionary process to generate human-like behaviours.