Agent technology has had an expanding impact and depth over the past two decades and more, influencing and being influenced by other disciplines while making a distinct research contribution with novel solutions to industrial problems. It is crucial, therefore, that both academics and industrialists in Europe have access to a forum at which current research and application issues are presented and discussed. The European Workshop on Multi-Agent Systems (EUMAS) series has played a distinctive role in encouraging and supporting activity in the research and development of multiagent systems, in particular bringing newer researchers into contact with experienced scholars and providing a venue for robust but friendly discussion on emerging ideas in the field. The overarching goals of the EUMAS workshop are to advance the state of the art in European research and development in the area of autonomous agents and multiagent systems, and to provide an environment for research training in this area. A nonprofit organization, the European Association for Multiagent Systems (EURAMAS), now coordinates the running of EUMAS, together with the European Agent Systems Summer School (EASSS).
EUMAS 2010 was held at Paris Descartes University (Paris, France) in December 2010, with two full days of two packed parallel sessions, preceded by another two days of co-located events: Workshops of the COST Action on Agreement Technologies, that year discussing Trust Technologies; and the Agent Technical Fora, providing focused half-day discussion forums on topics such as multiagent simulation and agent-oriented software engineering. EUMAS 2010 included talks by three prestigious invited speakers: Katia Sycara, Director of the Intelligent Software Agents Lab at Carnegie Mellon University; Sarit Kraus, Adjunct Professor of Computer Science at the University of Maryland and Professor of Computer Science at Bar Ilan University; and Jérôme Lang, senior researcher at Centre National de la Recherche Scientifique (CNRS).
Although EUMAS is deliberately informal, having no published proceedings, it still attracts a sizeable number of participants, reaching well beyond the borders of Europe. EUMAS 2010 included around 100 participants, with submitting authors from 23 different countries. Moreover, the submissions to EUMAS include a great deal of high-quality research, capturing the most novel areas in the field. This quality is evident in the selection of articles for this issue, which are revised and updated versions of a selection of EUMAS 2010 submissions considered particularly excellent by the workshop reviewers. The revised articles were reviewed by prominent researchers in the field specifically for this issue, with the following five being included.
Overall, the topics of the articles indicate development in some particular areas of multiagent systems research. Recurrent themes are the organizational structures, from loose coalitions to more strictly defined institutions, which influence agents' strategies and the emergent properties that those structures either cause or allow to be detected and manipulated.
First, in “CHARMS: A Charter Management System. Automating the Integration of Electronic Institutions and Humans,” Brito et al. consider the link between electronic institutions, which organize agents in defined processes by conventions defined in charters, and the users that specify, execute, and maintain those electronic institutions. Their CHARMS system generates graphical user interfaces with which users can connect to the institutions and adapts these interfaces when the institution specification changes, ensuring that the user can observe and provide input to the processes.
In “Model-Checking Alternating-Time Temporal Logic with Strategies Based on Common Knowledge Is Undecidable,” Diaconu and Dima study strategies of agents within coalitions sharing a pool of common knowledge. They present semantics for the Alternating-Time Temporal Logic (ATL) with imperfect information, and show that it is undecidable under certain conditions.
Next, Junges and Klügl, in “Programming Agent Behavior by Learning in Simulation Models,” apply multiple learning techniques to learning policies for agent behavior to achieve particular emergent, macro phenomena in multiagent simulations.
Also studying emergent phenomena, but in a quite different context, “A Decentralized Approach for Detecting Dynamically Changing Diffuse Event Sources in Noisy WSN Environments” by Fernandez-Marquez, Arcos, and Serugendo considers the detection of sources of diffuse events as they spread signals within an environment in which that detection relies on the observations and interactions of agents localized within the environment. They present a robust gradient-based approach to tracking these event sources.
Finally, Tampitsikas et al. present “Interdependent Artificial Institutions in Agent Environments,” in which they outline an approach to artificial institutions, governing agent behavior in terms of regulatory norms and powers. The novelty in their approach lies in how the institutions are specified, presenting them as first-class entities available to be inspected, manipulated, and modified by the agents themselves.
The success of the workshop and the quality of this issue are due to the voluntary efforts of many scholars who provided valuable feedback on the work in EUMAS 2010 reviews, in discussion at the workshop itself, and in preparing this special issue. We would particularly like to thank the following for their support in ensuring the high quality of this issue's articles (alphabetical order): Lina Barakat, Maria Chli, Marina De Vos, Michal Jakob, Wojtek Jamroga, Alessio Lomuscio, Pablo Noriega, Michael Schumacher, and Laurent Vercouter. We would further like to thank EURAMAS for allowing us to organize and host EUMAS in 2010, and to the editors of this journal for the opportunity to more widely disseminate the pick of the best submissions to the workshop.