Environments for Multi-Agent Systems II: Second International Workshop, E4MAS 2005, Utrecht, The Netherlands, July 25, 2005, Selected Revised and Invited Papers

International audienceIn the past two years the environment in multiagent systems has become increas-ingly important and is now becoming a focus of research in its own right. Yet,the environment in multiagent systems has been studied before. So the obviousquestion then is: Why does the environment attract the attention of a broadercommunity of researchers right now?The answer to this question is manifold. First, current research on environ-ments is built on the receptive ground of early work. Pioneers such as Demazeau,Parunak, Ferber, Odell, Omicini and Zambonelli have been stressing the impor-tance of the environment in multiagent systems for almost a decade.Second, current research on environments is well organized. The workshopseries on Environments for Multiagent Systems (E4MAS) provides the breedingground for coordinating research on environments. E4MAS provides an activeforum for discussion and exchange of ideas. The constructive atmosphere of theE4MAS workshops and the critical attitude of the attendees stimulate researchin the growing community.Third, researchers interested in environments come from various backgrounds.The notion of environment exceeds specific types of agency. The environment isimportant for simple ant-like agents as well as for complex cognitive agents. Theenvironment provides a challenging area for synergetic research on multiagentsystems in general.And last but not least, the perspective on the role of the environment inmultiagent systems has undergone a fundamental change in the last two years.Whereas environment and “infrastructure” are traditionally considered equiva-lent, recent research considers the environment as a first-order design abstractionin multiagent systems. Several researchers have demonstrated that the environ-ment provides a building block that can be used creatively in the design ofmultiagent system applications. Distinguishing between agent and environmentresponsibilities supports separation of concerns in multiagent systems, which isa prerequisite for good engineering practice.This volume collects selected and revised papers of the second E4MASWorkshop, completed with a number of invited papers of prominent researchersactive in the domain. The papers represent the full life-cycle of environmentengineering, including theoretical analysis, models, mechanisms, architecture anddesign, and applications. We hope that the work presented in this book stimulates further exploration and exploitation of the environment in multiagentsystems