Arguing for Computational Power

摘要

Nearly every book on agent-based models (including Generative Social Science) features advanced praise on the dust cover that highlights the tremendous promise of agent-based modeling. Agent-based modeling, however, is now at least 35 years old-counting from the publication of Thomas Schelling's Micromotives and Macrobehavior (1)-and its impact on mainstream research in the three major social sciences (economics, sociology, and political science) has been limited in contrast to, say, that of game theory. Before the reader jumps to conclusions about conspiracies among scientific elites that rigidly hold on to existing views, it is worth keeping in mind that over the past 15 years, the supposed fortress of orthodoxy, theoretical economics, has been revolutionized at least twice: first by the introduction of evolutionary game theory and more recently by the birth of behavioral economics and neuroeconomics. Agent-based models have had no similar impact.