About AAMAS
AAMAS is the leading scientific conference for research in autonomous agents and multiagent systems. The AAMAS conference series was initiated in 2002 by merging highly respected individual conferences:
– the International Conference on Multi-Agent Systems (ICMAS);
– the International Workshop on Agent Theories, Architectures, and Languages (ATAL);
– the International Conference on Autonomous Agents (AA).
The aim of the joint conference is to provide a single, high-profile, internationally respected archival forum for scientific research in the theory and practice of autonomous agents and multiagent systems. See http://www.ifaamas.org for more information.
AAMAS 2009 is the Eighth conference in the AAMAS series, following enormously successful previous conferences at Bologna, Italy (2002), Melbourne, Australia (2003), New York, USA (2004), Utrecht, The Netherlands (2005), Hakodate, Japan (2006), Honolulu, USA (2007) and Estoril, Portugal (2008). AAMAS-09 will be held at the Europa Congress Center, Budapest, Hungary. AAMAS 2009 is presented in cooperation with ECCAI, AAAI , ACM SIGART and the AI SIG of the Hungarian John von Neumann Society.
Introduction
In the beginning of the 21th century, the success of the so-called information society has changed dramatically both the requirements and the main characteristics of most computer-based systems. The previous century saw a migration from huge mainframes with limited capacity of memory to personal microcomputers that were further linked by powerful networks. Pushing one step further the evolution that computing has experienced since its beginning, today’s users are part of an on-line design team and can influence and change dynamically the contents and shape of information provided by computer-based systems. Moreover, computing is not restricted anymore to traditional desktop computers: with the parallel progress of telecommunications and information technology, we have plenty of applications running on mobile phones, PDAs, network interfaces and digital TV. These new scenarios impose complementary requirements for system design, such as ubiquity, distribution, and availability, generating thus an increase on these systems’ size, complexity, and diversity.
Agents and multiagent systems are a natural way to conceptualize, understand, analyze, design, and implement such systems, since from the beginning the area has aimed to deal with complex and distributed systems, composed of autonomous individual entities that were supposed to act, and interact, in an organized way.
Agent and multi-agent oriented models and technologies are being used in several domains of applied research, like social simulation, games, pervasive and ubiquitous computing, robotics, user interfaces, computer-mediated collaboration, electronic commerce, information retrieval, education and training. All of these research areas are represented in the AAMAS 2009 program.
The Autonomous Agents and Multiagent Systems (AAMAS) conference series brings together researchers from around the world to share the latest advances in the field. It was initiated in 2002 as a merger of three highly successful related events: the International Conference on Autonomous Agents (AGENTS), the International Conference on Multi-Agent Systems (ICMAS), and the International Workshop on Agent Theories, Architectures, and Languages (ATAL). The AAMAS conference series provides a single, high-profile forum for research in the theory and practice of autonomous agents and multiagent systems. AAMAS 2002, the first of the series, was held in Bologna, followed by AAMAS 2003 in Melbourne, AAMAS 2004 in New York, AAMAS 2005 in Utrecht, AAMAS 2006 in Hakodate, AAMAS 2007 in Honolulu, AAMAS 2008 in Estoril and AAMAS 2009 in Budapest.
The main theme of AAMAS-09, based on feedback from previous editions, is reinforcing the rich panorama of *interconnections* in the field.
We wanted to encourage the community, and the authors in particular, to reflect about their work not as belonging to just a niche in a long list of topics, but rather as a point in an abstract topological space defined by three complementary axes:
Focus axis
Broadly speaking, the AAMAS community-specific topics of interest are focussed in individual agents, their environment, their interactions, and their social/organizational structure.
• Agent Focus: contributions are focussed on the elements for structuring/defining/analyzing individual processing entities:
agent architectures
agent reasoning/deliberation/decision mechanisms
agent perception and action
believable agents
agents models of emotion, motivation, personality, etc.
• Environment Focus: contributions are focussed on the elements for structuring/defining/analyzing the external world (including humans) and its information exchange with the processing entities:
environmental entities
agent-human interaction
interface agents
environment architecture/structure
• Interaction Focus: contributions are focussed on the elements for structuring/defining/analyzing the information and control exchange between the processing entities
agent communication languages
interaction protocols
ontological / semantic interactions
conflict resolution and negotiation
argumentation theories
mechanism design
multi-agent learning
multi-agent planning
auctions and related structures
coordination mechanisms
reputation and trust
privacy and security
• Social/Organizational Structure Focus: contributions are focussed on elements for structuring/defining/analyzing the identity and properties of multiple processing entities
groups and teams
norms and normative behavior
commitments
autonomy
organizations and institutions
organizational learning
organizational planning
organizational reputation and trust
coalition formation
open systems
collective decision making
emergent behavior
• Comprehensive/Cross-cutting Focus: contributions dealing with all or more than one of the previous focus
complete applications demonstrating several different significant foci
agent-oriented software engineering (multiple foci)
agent and multi-agent programming languages (multiple foci)
multi-agent based simulation (multiple foci)
evaluation techniques
ethical and legal issues
standardization efforts in industry and commerce
meta-papers (on the state of agent research, on the conference, etc.)
Description level axis
Orthogonally to the focus axis, contributions to the field can be presented in several descriptive styles
• Theoretical, where the emphasis is in describing new theories of agents or agenthood using formal descriptions.
• Experimental / empirical, where the emphasis is on issues clearly in the context of autonomous agents or multi-agent systems architectures, simulations, or performance.
• Methodological / languages, where the emphasis is on agents and/or MAS analysis/design methodologies, verification and validation techniques/tools, and multi-agent programming languages, deployment platforms and development tools.
• Applications, where the emphasis is in the motivation for an agent / MAS approach, and a clear description of the design and implementation / evaluation of the system.
Inspiration source axis
This axis is related to the different scientific domains that have provided the main ideas for a contribution. AAMAS-09 welcomed papers from many intellectual approaches, including artificial intelligence, distributed systems, economics, social/management sciences, (multi-)robotics, and biologically-inspired approaches.
As an example, a paper describing formally an extension of a BDI theory would be classified as (agent, theoretical, artificial intelligence), whereas the description of an interpreter of this extension could be classified as (agent, methodological/languages, artificial intelligence) and an application for several agents interacting by using this extension could be mapped as (comprehensive, applications, artificial intelligence). Hence, this topological space could help this year and future AAMAS conference organizers to better organize the presentation sessions and to stress some particular points in this space to be highlighted in the Conference program in a more principled way.
AAMAS 2009 has encouraged papers that focus on some of the issues located in the topological space described above:
Inspiration source axis:
(Multi-)Robotics (following the 2008 Special Track)
Social Sciences
Description level axis:
Methodologies/languages
Applications
Focus axis:
(Virtual) Agents (following the 2008 Special Track)
Social/Organizational Structure
This year the conference solicited regular papers (8 pages) and extended abstracts (2 pages). Extended abstracts were encouraged as a mechanism for the timely reporting of interesting but preliminary work, that may not as yet have the level of evaluation or detail that would be expected for a regular paper. The program chairs could, at their discretion, accept papers that were submitted as regular papers as extended abstracts, if the authors explicitly agreed to this when registering their papers.
Accepted regular papers received a slot for oral presentation in the conference. In addition, both regular and extended abstracts were presented in poster sessions and appeared in the conference Proceedings.
Reviews were double blind, as authors were requested to avoid including anything that could be used to reveal their identity. Submissions were peer reviewed rigorously and evaluated on the basis of originality, soundness, significance, presentation, understanding of the state of the art, as well as the overall quality of their technical contribution. Each paper was reviewed by at least three Program Committee members, coordinated by a Senior Program Committee member.
Another innovation was the establishment of Area Chairs, which were responsible for having a general view on some particular aspect of the classification axes, and who helped the Program Chairs to produce a more balanced program.
AAMAS 2009 received a total of 651 submissions for the main conference, divided between 591 (91%) full paper submissions and 60 (9%) extended abstract submissions. These contributions came from 21 countries: 380 came from Europe, 202 from North America. and 69 from the rest of the world.
As a result of the reviewing procedure, 132 contributions were accepted as full papers (an acceptance rate of 22 %). and 128 as extended abstracts. Within these latter, 9 were originally submitted as extended abstracts (an acceptance rate of 15 %) and the other 119 (20 %) were submitted as full papers, but accepted as extended abstracts.
AAMAS prides itself on being a nurturing environment for students working in the area. 60 % of the submitted full papers and 52 % of the submitted extended abstracts had a student as the primary author, with a respective acceptance rate of 23 % for full papers, 16 % for extended abstracts (submitted as such), and 21 % for extended abstracts, though submitted as full papers. These numbers reveal that the statistics of submission/ acceptance of students as first authors are very similar to the global statistics of the Conference.
The submissions were divided in the following way, according to our abstract classification space:
Description Level
Applications 8%
Experimental/Empirical 43%
Theoretical 32%
Methodological/Languages 17%
Inspiration Source
AI 52%
Biological Approaches 2%
Distributed Systems 14%
Economics 11%
Multi-Robotics 5%
Social/Management Science 10%
Other 6%
Focus
Agents 24%
Environments 4%
Interactions 39%
Social/Organizational Structure 17%
Cross-cutting 16%
The special tracks have also attracted a great number of submissions:
Tracks:
[Inspiration]
(Multi-)Robotics: 31 papers
Social Sciences: 68 papers
[Description Level]
Methodologies/ Languages: 111 papers
Applications: 55 papers
[Focus]
(Virtual) Agents: 155 papers
Specifically Believability, Emotion/Personality, or Human/Interaction: 34
Social/Organizational Structure: 109 papers
The main conference includes two invited talks of well-known researchers in the field:
• Michael N. Huhns (University of South Carolina, USA)
“From DPS to MAS to ... Continuing the Trends”
• Klaus G. Troitzsch
(Universitat Koblenz-Landau, Germany)
“Perspectives and Challenges of Agent-Based Simulation as a Tool for Economics and Other Social Sciences”
Moreover, two panels that address important issues for the MAS community are expected to happen:
• Agent Oriented Methodologies and Programming Languages: Towards Practical Systems
The main goal of this panel is to present the state of the art of agent oriented software engineering (AOSE) methodologies, agent oriented programming languages (AOL) and agent oriented applications (AOA). Moreover, the idea is to discuss to what extent these complementary issues are interrelated. This panel has as participants Klaus Fischer (DKKI, Germany), Andrea Omicini (University of Bologna, Italy), Amal El Fallah Seghrouchni (University Paris VI, France) and Milind Tambe (University of Southern California, USA).
• Theoretical Foundations for Agents and MAS: Is Game Theory Sufficient?
The main goal of this panel is to discuss the importance of game theory (GT) as a fundamental theoretical basis for MAS. Even if there is no doubt that GT originated a significant number of MAS advances, there are some authors that argue that GT can not capture in a proper abstraction level all the different phenomena involved in autonomous agents reasoning and decision mechanisms, MAS interactions and MAS organizations. Moreover, the idea is to discuss to what extent other theoretical foundations can be used to address these phenomena and how they are/ could be interrelated with GT. This panel has as participants Rosaria Conte (ISTC, Italy), Jacques Ferber (University of Montpellier, France), Wiebe van der Hoek (University of Liverpool, UK) and Jeffrey Rosenschein (Hebrew University of Jerusalem, Israel).
Besides the main conference program, the six-day AAMAS event consists of three days of workshops and tutorials, with a total of 9 tutorials and 26 workshops. The three day main conference includes a separate industry track with 10 full and 2 short papers, demonstration sessions, and a small exhibition. A doctoral consortium precedes the workshops and tutorials. These and other conference elements are the result of great efforts on the part of many dedicated volunteers. We are grateful to all the AAMAS 2009 conference officials and those who worked with them to put together such a successful research meeting.
The chairs would like to thank all of the Program Committee, Senior Program Committee and Area Chair members for their hard work putting together this year’s program. Overall, the reviews were quite detailed and the reviewers were very responsive during the discussion phase.
We are especially grateful to Lin Padgham, David Parkes, and Jomi Fred Hübner, for having respectively developed and customized the paper allocation system that we have used this year, Viktor Richer and Gusztav Hencsey for their hard work in assembling this proceedings, and to Eva Thiry for designing and maintaining the Conference website.
Jaime Sichman and Keith Decker,
AAMAS 09 Program co-Chairs
Carles Sierra and Cristiano Castelfranchi,
AAMAS 09 General co-Chairs.