Multi-Agent-Based Simulation VIII

This volume contains selected papers that were presented at the eighth int- national workshop on Multi-Agent-Based Simulation (MABS 2007), a workshop co-located with the 6th International Joint Conference on Autonomous Agents andMulti-Agent Systems (AAMAS 2007),held inHonolulu, Hawaii,onMay15, 2007. These papers have been revised and extended, based on discussions at the workshop, and reviewed once more. Agenttechnology is now a mature paradigm of software engineering. C- plex systems, which are irreducible to their components in isolation, are instead heavily characterizedby the interaction between their components. Agent-based simulation is the natural way to model systems with a focus on interaction, and the circle closes by considering how the social sciences show this kind of c- plexity. The focus of this workshopseries lies in this con?uence of socialsciences and multi-agent systems. 1 Simulation has been proposed by Axelrod as athirdwayofdoingscience, in contrast with deduction and induction: generating data that can be analyzed inductively, but coming from a rigourously speci?ed set of rules rather than - rect measurement of the real world. In this sense, to simulate a phenomenon is to generate it - constructing arti?cial (agent) societies. This in turn leads to questions that have already been asked for human societies. Computer sci- tists have adopted general terms like emerging behavior, self-organization, and evolutionary theory; even speci?c social terms such as norms, reputation, trust, tags, institutions; but all of them in an intuitive manner.