Agent-Based Models of Social Intelligence Based on Behavioral Heuristics Informed by Cognitive Neurosciences

The broader objective is to develop a new perspective or modeling economic interactions that goes beyond the standard paradigm of Homo economicus, in which the recent financial crisis has demonstrated important efficiencies.

Agent-based models have been used extensively in the computer science and economics literatures to generate macroscopic implications of learning behavior, competition, and evolution in systems too complex to solve analytically. By positing relatively simple behavioral heuristics and simulating the random interactions of many agents with such heuristics, surprising aggregate phenomena sometimes emerge, e.g., limit cycles, phase transitions, and market crashes. However, the practical relevance of such findings is questionable because of the simplistic and often counter-factual behavioral heuristics assumed. Our broader objective is to develop a new perspective or modeling economic interactions that goes beyond the standard paradigm of Homo economicus, in which the recent financial crisis has demonstrated important efficiencies. This new perspective will be particularly important in changing the way we think of regulation and public policy, both of which are forms of social intelligence and whose functions can be significantly improved by adopting a more realistic framework for economic behavior.