The South Eastern European
Multi-Agent Systems
Summer School
Bucharest, Romania
July 5-10, 2010


Venue & registration: July 4, 2010
School: July 5-9, 2010
Trip to Bucegi Mountains: July 10, 2010


Mon 5 Tue 6 Wed 7 Thu 8 Fri 9 Sat 10
Student session
Student session
Conclusions and round table
Student session
Round table
Social programme



Michael Luck, King’s College London

COST AT Lecturer

Michael Luck is Professor of Computer Science in the Department of Computer Science at King’s College London, where he leads the Agents and Intelligent Systems group and undertakes research into agent technologies and intelligent systems. His work has sought to take a principled approach to the development of practical agent systems, and spans formal models and theories as well as practical applications. Recent work has been directed at norms and institutions, declarative programming of agent systems, and industrial deployment and technology forecasting. He is currently leading work at King’s on the IST CONTRACT project, concerned with distributed electronic business systems on the basis of dynamically generated, cross-organisational contracts.

Professor Luck has published around 200 articles in these and related areas, and twelve books (including monographs, textbooks, and edited collections); he was lead author of the AgentLink roadmaps in 2003 and 2005. He is a a director of the board of the International Foundation for Autonomous Agents and Multi-Agent Systems (IFAAMAS), co-founder of the European Multi-Agent Systems (EUMAS) workshop series, co-founder and Chair of the steering committee of the UK Multi-Agent Systems Workshops (UKMAS), and a Steering Committee member for the Central and Eastern European Conference on Multi-Agent Systems (CEEMAS). Professor Luck was a member of the Executive Committee of AgentLink III, the European Network of Excellence for Agent-Based Computing, having previously been the Director of AgentLink II. He is an editorial board member of Autonomous Agents and Multi-Agent Systems, the International Journal of Agent-Oriented Software Engineering, Web Intelligence and Agent Systems, and ACM Transactions on Autonomous and Adaptive Systems, and was previously series editor for Artech House’s Agent Oriented Systems book series. Michael Luck is general co-chair of the Ninth International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2010), to be held in Toronto, Canada in May 2010.

Introduction to Multi-agent Systems

In this talk I will provide an introduction to intelligent agents and multi-agent systems, covering the motivation for the use of agents, BDI concepts, agent interaction, applications and future prospects.

John Jules Meyer, Utrecht University, The Netherlands

COST AT Lecturer

Prof.dr. John-Jules Ch. Meyer studied mathematics with computer science and digital signal processing at Leyden University. In 1985 he obtained his Ph.D. from the Vrije Universiteit in Amsterdam on a thesis entitled “Programming Calculi Based on Fixed Point Transformations”, a subject in theoretical computer science. From 1988 to 1993 he was a professor at the computer science department at the VU Amsterdam holding a chair in “Logic for distributed systems and artificial intelligence”. From 1989 to 1993 he also was a professor of theoretical computer science at the Katholieke Universiteit Nijmegen. Since 1993 he has been a professor at the computer science department of Utrecht University (UU). At the moment he is heading the Intelligent Systems Group of the Department of Computing and information Sciences of the UU.

In 2007-2008 he was a guest professor/researcher at IDA, Linkoeping University. He is a member of the IFAAMAS board steering the international AAMAS conferences, and of the editorial boards of the Journal of Applied Non-Classical Logics, Data and Knowledge Engineering, the Journal of Intelligent Agents & Multi-Agent Systems, and the International Journal of Reasoning-Based Intelligent Systems. In 2005 he was appointed as a Fellow of the European Coordinating Committee for Artificial Intelligence. He has been included in The International Directory of Logicians: Who’s Who in Logic in 2009. Currently he is also general and scientific director of the Turing Institute Almere, which specializes in multi-disciplinary (including AI and agent) research for the health sciences, and senior advisor at TNO Human Factors, Soesterberg. His current research interests include logics for AI, intelligent agents and cognitive robotics, and he has been involved with agent research for over 15 years now, ranging from theoretical / logical foundations, via agent programming to the various applications that are currently being investigated in his group.

Logical Foundations of Multi-agent Systems

In this tutorial I will treat the basics of modal logic, and how they are applied to describe agent attitudes and multi-agent systems. Topics that will be addressed are basic modal logic, epistemic logic, temporal logic, dynamic logic, deontic logic as well as BDI logic, Common Knowledge and Joint Intentions, time permitting. I will end with non-normal modal logic (neighborhood semantics), which provides the basis for MAS logics such as Coalition Logic.

Wojciech Jamroga, University of Luxemburg, Grand Duchy of Luxembourg

Wojciech Jamroga did his PhD at the University of Twente (the Netherlands), and his habilitation at the Clausthal University of Technology (Germany). Currently he is a member of the Individual and Collective Reasoning Group at the University of Luxembourg. He works mainly on logical and game-theoretical aspects of multi-agent systems.

His teaching record includes four courses at ESSLLI (European Summer School on Logic, Language and Information: 2006, 2007, 2008, and forthcoming in 2010), and four courses at EASSS (European Agent Systems Summer School: 2006, 2007, 2008, and 2009).

Reasoning about Strategies of Agents

A multi-agent system can be seen as a game where agents act to obtain a desirable outcome. Thus, when reasoning about such systems, agents’ strategies (and their outcomes) must be taken into account. On one hand, these issues have been studied extensively in game theory. On the other hand, formal logic is widely regarded as a foundation for specification, verification and reasoning about MAS. In this course, we study logical formalisations of agents in game-like scenarios.

The course is focused on Alternating-time Temporal Logic (ATL), one of the most prominent logics of cooperation and strategic ability. We introduce ATL, show its relationship to extensive games of perfect information, and discuss some important extensions of the logic, e.g. with motivational attitudes of agents and/or rationality assumptions. In particular, we focus on imperfect information and the integration of strategic and epistemic modalities. We also discuss the verification problem for ATL, and give an overview of complexity results.

Sarit Kraus, University of Maryland, US and Bar Ilan University, Israel

Sarit Kraus (Ph.D. Computer Science, Hebrew University, 1989) is a Professor of Computer Science at Bar-Ilan University and Adjunct Professor at the Institute for Advanced Computer Studies, University of Maryland (UMIACS). Her research interests are in multi-agent systems, specially negotiation and cooperation among agents, open agent environments, learning and information agents, personalization and optimization of complex systems.

In 1995 Kraus was awarded the IJCAI Computers and Thought Award (the premier award for a young AI scientist). In 2001 she was awarded the IBM Faculty Partnership Award and in 2002 she was elected as AAAI fellow. In 2007 she was awarded the ACM SIGART Agents Research award and her paper with Prof. Barbara Grosz was a winner of the IFAAMAS influential paper award (joint winner). In 2008 she was elected as ECCAI fellow. She has published over 250 papers in leading journals and major conferences and is an author of the book Strategic Negotiation in Multiagent Environments(2001) and a co-author of a book on Heterogeneous Active Agents (2000); both published in MIT Press. She served as an associate editor of the Artificial Intelligence journal during 2004- 2008 and since 2002 she is the Senior Associate Editor of the Annals of Mathematics and Artificial Intelligence.

Home page:

Automated negotiators: Agents interacting with other automated agents and with humans
Negotiation is an important mechanism for resolving conflicts and is prevalent in human interaction in many different settings, such as diplomatic relations, electronic commerce, and personal relationships. In this tutorial we will present the key approaches for designing automated negotiators. These include game theoretic and heuristics and their integration with machine learning methods. Game theory provides an elegant mathematical framework for modeling and analyzing strategic interaction between self-interested fully rational agents. However, implementation of these models for agents that need to interact with people is problematic since people are rationally bounded and affected by social factors.

This tutorial will present a variety of communication protocols (e.g., alternating offers, flexible commitments), decision-making strategies (e.g., equilibrium solutions, opponent modeling), and mechanisms (auctions, voting) for achieving efficient negotiation. We will combine theoretical and algorithm results with empirical frameworks designed to facilitate the study of negotiation in environments that include people and computer agents. The tutorial does not assume prior background in game theory.

Human-Computer Negotiation: Learning from Different Cultures

Negotiation is a process by which interested parties confer with the aim of reaching agreements. The ability to negotiate successfully is critical for many social interactions. The dissemination of applications such as the Internet across geographical and ethnic borders are opening up opportunities for computer agents to negotiate with people of diverse cultural and organizational affiliation. These automated negotiators should be able to proficiently interact and collaborate with their human partners.

However, people often follow suboptimal decision strategies due to irrationalities attributed to lack of knowledge of own preferences, the effects of the task complexity, framing effects, the interplay between emotion and cognition, and the problem of self control. Furthermore, culture plays an important role in their decision making and people of varying cultures differ in the way make offers and fulfill their commitments in negotiation. In this talk we will describe a new agent for bilateral negotiation in repeated interactions that allow players to renege on agreements. PURB was especially designed to adapt to the particular behavioural traits of its negotiation partner. Its strategy is composed of a utility function that depends on the extent to which the other player is reliable and helpful, as well as rule-based mechanism that uses this utility for generating and replying to offers, and for deciding whether to fulfill its agreements. I will present an extensive study that evaluated the performance of PURB when interacting with people in three different cultures. Our results show that the performance of the PURB agent directly depended on the cultural affiliation of its negotiation partners: People’s negotiation behaviour, in particular the extent to which they fulfill agreements, varies widely across cultures, and this had a crucial effect on PURB’s performance. I will also present additional results that compare the performance of PURB when playing Peer-Designed-Agents — agents that were designed by non-experts to represent themselves during negotiation. I will conclude by demonstrating how the PURB agent could be improved using the collected data.

Amal El Fallah Seghrouchni, University Pierre et Marie Curie, France

Amal El Fallah Seghrouchni is Full Professor at the University Pierre and Marie Curie (Paris – France). She is researcher at LIP6 laboratory (University Pierre and Marie Curie and CNRS – UMR7606) where she heads the Multi-Agent Systems team. She is also the president of the French “Collège SMA” which federates French teams working in the domain of intelligent agents and Multi-Agent Systems.

Her main research topics are “Analysis, Design and Validation of Multi-Agent Systems”. She published several papers related to MAS field, co-edited books or post-proceedings and co-organised several international events dedicated to Multi-Agent Systems (e.g. Dagstuhl seminars ; the series of ProMAS workshops as a satellite event of AAMAS since 2003, and more LADS workshop at MALLOW since 2007). She also has been invited professor and gave talks in international seminar and courses for up-graduated students about multi-agent planning, coordination and interaction protocols, mobile computing, etc. (for more details, please visit http:///

Multi-Agent Planning and coordination in MAS
One of the major interests of Multi-Agent Systems (MAS) is their ability to handle distributed planning by coordinating agents’ plans. Coordination requires both an adequate plan representation and an efficient interaction between agents. Based on information exchange (e.g. data, plans), the interaction allows agents to update their own plans by considering the exchanged information. Coordination generally produces two effects: cancelling negative interactions (e.g. harmful actions) and taking advantage of helpful interactions (e.g. handling redundant actions). Agents organize their activities and update their plans in order to cooperate and avoid conflicts.

This lecture will address multi-agent planning as an important mechanism for MAS coordination. From the theoretical point of view it raises several interesting issues because of the main features of agents (autonomy and partial view of agents) and the environment changes. Multi-agent planning will be studied from several perspectives: dynamic or reactive planning, distributed planning, task allocation and resource sharing, etc.

Programming principles of MAS

Intelligent agents and multi-agent systems (MAS) play an important role in today’s software development. Indeed, they constitute an interesting paradigm to build distributed intelligent systems and offer a relevant abstraction to design complex applications at industrial level.
Since almost two decades, the MAS community has developed a large and rich panel of concepts, architectures, interaction techniques, and general approaches to the analysis and the specification of MAS. The aim of this presentation is to survey the main principles of MAS programming and to highlight some of their links with (distributed) artificial intelligence concepts, distributed systems techniques; and software engineering methodologies.

Eugénio Oliveira, Universidade do Porto, Portugal

COST AT Lecturer

Eugénio Oliveira studied at the Universities of Coimbra and Porto in Portugal where he got his degree in Electrotechnical Engineering in the seventies. He immediately went to Switzerland, near Zurich, where he was enrolled as R&D Engineer at Brown, Boveri, C. (now ABB) Electronic Labs. He got his PhD at the New University of Lisbon with a thesis on Knowledge Engineering and Logic Programming applications (1984). Eugénio Oliveira was invited as a Guest Academic at IBM IEC in La Hulpe, near Brussels (84-85) teaching Knowledge Based Systems and launching projects on Logic Programming and Expert Systems. He started, in 987, the Distributed Artificial Intelligence & Robotics Group at the Faculty of Engineering of the University of Porto and, in the next year, he was co-founder of LIACC- AI and Computer Science Lab at the University of Porto. LIACC has about 60 members in 2010. LIACC includes groups working on Multi-Agent Systems, Intelligent Robotics and Computer Science.

Eugénio Oliveira, besides chairing both DAIR group and LIACC Lab (2007-10), is a Full Professor in Informatics at the University of Porto since 2000. He got his “Habilitation” title in 1991 at the University of Porto. He is Director of the PhD Program on Informatics Engineering. He was appointed as Director for the creation of the Informatics Department at the Faculty of Engineering, University of Porto (2007-08). He successfully supervised 10 PhD students and more than 40 Master students. He was involved in many research projects, some of which funded by the European Union as it was the case of the ARCHON project, a five years main project on agent applications in the 90s. He was awarded the Gulbenkian Prize for Science and Technology in 1983. He published more than 150 papers, mainly in the field of software agents and multi-agent system applications as well as intelligent robotics. He has recently published on text mining, agent-based electronic institutions and Trust & Reputation models. He was one of the area chairs at ECAI 94, ECAI 96 and ECAI 2008. He was involved in AGENTLink and is Portuguese delegate and co-founder of “Agreement Technology” COST Action (IC0801) funded by EU (since 2008). He is member of the Editorial Boards of the Journals: “Autonomous Agents and Multi-Agents Systems International Journal”, “Agent Oriented Software Engineering” and “Intelligent Decision Technologies”.

Agent Advanced Features and Electronic Institutions for B2B Interoperability

Electronic Institutions (EI) can provide a good environment as well as appropriate services to facilitate B2B Interoperability.
In my tutorial I will elaborate on different aspects that have to be present, through the EI, to make easier Virtual Enterprises formation and operation stages. I will elaborate on how to design a Normative Environment to help on electronic contract monitoring, how to provide an adaptive negotiation protocol and on Computational Trust and Reputation Models that may help in the partners’ selection operation.

Laurent Vercouter, École Nationale Supérieure des Mines de Saint Etienne, France

Dr. Laurent Vercouter is an assistant professor and a member of the multiagent systems research department at the Ecole des Mines of Saint-Etienne in France. The global research theme of L. Vercouter is decentralized multiagent systems and more specifically trust and reputation in MAS. In this field he developed in collaboration with Guillaume Muller a reputation model for the detection and exclusion of liars in P2P systems and published several papers internationally. In 2006, he worked as an invited researcher at the University of Sao Paulo on ontological aspects of reputation.

He is involved in the national project ForTrust about the formalization of trust for multiagent systems and a funding member of the international ART Testbed group.

Computational Trust and Reputation Models

The talk will present recent existing works in trust and reputation systems for multi-agent systems. First, the typical problems tackled by trust and reputation models are presented. We show that they are different problems that the ones tackled by security techniques and that security and trust must be considered as complementary approaches. The foundations of trust models are then explained. Starting from sociological studies, we identify the main concepts involved in trust models. Then we will survey the main existing trust models in MAS covering both the socio-cognitive and the game theorical approaches. Reputation systems will also be presented to show how they can contribute to (semi-)automatic trust decision processes. Finally, the talk will be concluded with an overview of current open research problems of the field.

Maria Ganzha, Marcin Paprzycki, Polish Academy of Sciences, Poland

COST AT Lecturer

Marcin Paprzycki (Senior Member of the IEEE, Senior Member of the ACM, Senior Fulbright Lecturer, IEEE CS Distinguished Visitor) has received his M.S. Degree in 1986 from Adam Mickiewicz University in Poznań, Poland, his Ph.D. in 1990 from Southern Methodist University in Dallas, Texas and his Doctor of Science Degree from Bulgarian Academy of Sciences in 2008. His initial research interests were in high performance computing and parallel computing, high performance linear algebra in particular. Over time they evolved toward distributed systems and Internet-based computing; in particular, agent systems. He has published more than 200 research papers and was invited to Program Committees of over 300 international conferences. He is on editorial boards of 14 journals and a book series.

Maria Ganzha obtained MS and her Ph.D. in Applied Mathematics from Moscow State University, Moscow, Russia in 1987 and 1991 respectively. Her initial research interests were in the area ofdifferential equations, solving mixed wave equations in space with disappearing obstacles in particular, currently she works in the areas of software engineering, distributed computing and agent systems in particular. She has published more than 70 research papers and is on editorial boards of 5 journals and a book series and was invited to Program Committees of over 40 conferences.

Software Agents as Resource Brokers in Grid

It is a widely held belief that software agents will become the next revolution in information technology. One of the areas where they are expected to play an important role is the Grid. Claims to this effect, as well as research results can be found in work of B. diMartino, O. Rana, B. Prasad and others. In our work we have taken a different approach to these researchers and proposed that agent teams should be utilized for resource brokering and management. In this way, software agents become the “brain” for the Grid “brawn.” The aim of the presentation will be to outline the assumptions underlying our work, introduce our system and discuss how agents representing users interact with the agent teams, and how the high-level intelligent infrastructure can utilize the actual Grid middleware to execute a job.

Social Programme

July 10 – Trip to Bucegi Mountains