Agent 2003 Conference on Challenges in Social Simulation

Welcome to the Proceedings of the fourth in a series of agent simulation conferences cosponsored by Argonne National Laboratory and The University of Chicago. Agent 2003 is the second conference in which three Special Interest Groups from the North American Association for Computational Social and Organizational Science (NAACSOS) have been involved in planning the program--Computational Social Theory; Simulation Applications; and Methods, Toolkits and Techniques. The theme of Agent 2003, Challenges in Social Simulation, is especially relevant, as there seems to be no shortage of such challenges. Agent simulation has been applied with increasing frequency to social domains for several decades, and its promise is clear and increasingly visible. Like any nascent scientific methodology, however, it faces a number of problems or issues that must be addressed in order to progress. These challenges include: (1) Validating models relative to the social settings they are designed to represent; (2) Developing agents and interactions simple enough to understand but sufficiently complex to do justice to the social processes of interest; (3) Bridging the gap between empirically spare artificial societies and naturally occurring social phenomena; (4) Building multi-level models that span processes across domains; (5) Promoting a dialog among theoretical, qualitative, and empirical socialmore » scientists and area experts, on the one hand, and mathematical and computational modelers and engineers, on the other; (6) Using that dialog to facilitate substantive progress in the social sciences; and (7) Fulfilling the aspirations of users in business, government, and other application areas, while recognizing and addressing the preceding challenges. Although this list hardly exhausts the challenges the field faces, it does identify topics addressed throughout the presentations of Agent 2003. Agent 2003 is part of a much larger process in which new methods and techniques are applied to difficult social issues. Among the resources that give us the prospect of success is the innovative and transdisciplinary research community being built. We believe that Agent 2003 contributes to further progress in computational modeling of social processes, and we hope that you find these Proceedings to be stimulating and rewarding. As the horizons of this transdiscipline continue to emerge and converge, we hope to provide similar forums that will promote development of agent simulation modeling in the years to come.« less