Analyzing the ENRON Communication Network Using Agent-Based Simulation

Abstract

Agent-based modeling, simulation, and network analysis approaches are one of the emergent techniques among soft computing literature. This paper presents an agent-based model for analyzing the characteristics of peer-to-peer human communication networks. We focus on the process of the collapse of Enron Corporation, which is an interesting topic among the business management domain. The Enron email dataset is available for the analysis. Our approach consists of the four steps: First, macro-level characteristics of the Enron email dataset is analyzed from the viewpoints of social network theory: (i) the degrees of the communication networks and contents information, and (ii) the changes of network structures among the major events. Second, for the micro-level analysis, an agent-based simulator is implemented using the Enron email dataset. Third, both micro- and macro- level characteristics are calculated on the simulator to ground the model to the dataset. Finally, a different artificial society from the Enron email dataset is developed the simulator and we compare its characteristics of communication patterns with the result of the ones in the agent-based simulation with the Enron email dataset. The investigation suggests that the agent-based model is beneficial to uncover the characteristics of implicit communication mechanisms of the firm.