Coordination Dynamics under Collective and Random Fining Systems for Controlling Non-Point Source Pollution: A Simulation Approach with Genetic Algorithms

The paper considers the application of Genetic Algorithms (GA) on coordination games with non-point pollution controlling systems including collective and random fines. During the GA simulations populations continually switch between the two symmetric Nash equilibria of the game. Coordination of GA populations on the socially optimal (payoff dominant) abatement equilibrium is observable even for large group sizes, if the fine is chosen high enough compared to the abatement cost savings in the zero-abatement (risk dominant) equilibrium. The time, which populations spend in the socially optimal equilibrium, declines strongly with increasing group size. The outcome of the random fining mechanism depends crucially on the underlying risk attitude. Purely risk averse populations tend to coordinate on the socially optimal equilibrium. Further, we examine whether the findings of the GA application can explain some main results from a series of related experiments, which are currently conducted by the Laboratory of Experimental Economics at the University of Ki