People

Publications

Projects

Software

News

We aim to develop a radically new class of tools with which to analyze the emergent behavior of systems for agent-based computing. Our research is based on a common theme in the history of science — new tools for representation and analysis of data often spur fundamental advances. We believe that research in agent-based computing will accelerate if we can provide investigators with better methods to record and analyze the behavior of their agent-based systems. Investigators will gain greater insight into intentional behaviors of their agents, clearer descriptions of behaviors that emerge from interactions among many agents, and faster identification of pathological behaviors. Perhaps most importantly, they will be better equipped to characterize the fundamental principles by which their systems operate, laying the foundation for a science of agent-based computing.

Our analysis techniques and software tools are built on a foundation of analytic approaches developed within intelligence analysis and social network analysis. These approaches are almost unknown in computer science, but we believe that they are uniquely suited to help understand the behavior of agent-based computing systems. Specifically, we are exploring relational data representations that are common in intelligence analysis and quantitative sociology, but rarely used in statistics and data mining. We conjecture that such representations can capture many of the important aspects of the behavior of agent-based systems, and that they can capture far more of those behaviors than traditional representations. Our primary aim to develop new tools and techniques that facilitate analysis of data recorded in relational representations.