Agent-based simulation for policy issue

The purpose of the paper is to analyze the role of a tax and a subsidy through the government for solving the social dilemma in a society of agents. We consider the situation where a group of agents is repeatedly matched to play a prisoner's dilemma game. Each agent only interacts with his neighbors, and when agents react, they react myopically (the myopia hypothesis). Agents are completely naive and do not perform optimization calculations. Rather, agents sometimes observe the current performance of other agents, and simply mimic the most successful strategy. The government levies taxes and provides them to agents. We show the role of the government is essential for designing a mutual reliable society.