The missing legacy of herbert simon in agent-based computational economics

Abstract

In this article, we examine the legacy of Simon in Agent-Based Computational Economics (ACE). We show that both near decomposability and modularity, the two essential ingredients of the Simonian economics, have not been seriously pursued by the ACE community. First, most ACE models are not endogenously multi-level, which makes near decomposability be not much relevant to ACE. Second, while the modularity approach has already been employed by Simon in his artificial intelligence research and can help shape a notion of autonomous agents, this approach is also not well followed by the ACE community. Instead, most artificial agents used in ACE have been put in rather identically repeated environment not much different what the movie "Groundhog Day" depicts [23]. Hence, they are not able to do serious novelty or chance discovery, and the creativity of artificial agents have not been taken seriously by most ACEers.