Rewarding and punishing evolving collusion in a simulation of an oligopolistic marketplace

The time phased behaviour of participants with interdependent pricing strategies in a marketplace is simulated using a GA. Each member of the population embodies a pricing strategy and their fitness equates to their profitability. Population members are monitored for collusive behaviour which is illegal in many real marketplaces because it is perceived to unfairly boost the profits of the colluders. In many of these same marketplaces a neutral authority has the power to punish collusion when detected. The marketplace simulation incorporates both the benefits and rewards of collusion. An investigation of the effects of varying the relative magnitudes of benefits and rewards and the time lag between the reaping of rewards and imposition of penalties has been made.