|
Motivation.
Advances in networking technology have revitalized
the investigation of the agent technology as a promising paradigm to engineer
complex distributed software systems. Nowadays, the agent technology has been
applied in a wide range of application domains, including e-commerce,
human-computer interfaces, telecommunications, and concurrent engineering. In
general, software agents are viewed as complex objects with an attitude. Like
objects, agents provide a specific set of services for their users. In fact,
objects and agents exhibit points of similarity, but the development of
multi-agent systems (MASs) poses other challenges to software engineering since
software agents are inherently more complex abstractions. A single agent is
driven by beliefs, goals, plans, and a number of behavioral properties such as
autonomy, adaptation, interaction, collaboration, learning and mobility. Each of
these features introduces additional complexity to the system modeling, design
and implementation, and consequently, increases the probability of exceptional
situation manifestation, security violations and so on. In addition, as the
agent paradigm is devoted to the complex distributed system development, a
large-scale MAS encompasses multiple types of agents, each of them having
distinct agency properties, and it needs to satisfy multiple stringent
requirements such as reliability, security, adaptability, interoperability,
scalability, maintainability, and reusability. However, many existing
agent-oriented solutions are far from ideal; in practice, they are often built
in an ad-hoc manner and are error-prone, not scalable, and not generally
applicable to large MAS. This workshop, aimed to discuss these issues, follows
up on the success of the 1st Workshop on Software Engineering for Large-Scale
Multi-Agent Systems (SELMAS’02), held in Orlando, USA, May 19, 2002, as part of
the ICSE’02. The 1st SELMAS workshop was very successful due to the high quality
of the submissions, the active participation of the audience, the exceptional
profile of the panelists, and the current edition of a Springer book publishing
refereed workshop papers.
Current Software Engineering Research.
Research in multi-agent software engineering has
been carried out according two different approaches: (i) agent-based software
engineering, and (ii) object-oriented software engineering for MASs. Researchers
following the first approach persuasively argue that multi-agent systems are
often much more complex than object-oriented systems and hence the traditional
object model generally fails to capture the complexity of multi-agent systems.
In this approach, agents are a new abstraction that substitutes for the object
abstraction realizing the agent abstraction as a new software engineering
paradigm. As a result, proponents of this approach claim that it is necessary to
develop new software engineering techniques, methods, and methodologies that are
specifically tailored to agents. On the other hand, researchers adhering the
second approach propose the integration of agents into the object-orientation
world and, thus, they think of objects and agents as similar abstractions. As a
result, their research has focused on using and extending the techniques
existing in object-oriented software engineering, such as design patterns,
frameworks, and modeling languages, to multi-agent software engineering.
Goals.
The goal of this workshop is to bring together researchers and practitioners
from both approaches to discuss the current state and future direction of
research in software engineering for large-scale MASs. In addition, the workshop
is a forum to learn about the latest research, and also discuss and exchange
ideas concerning ongoing work. Particular interests of this workshop are:
-
Determine
the overlap and integration of the two general research approaches for
multi-agent software engineering;
-
Understand
those issues in the agent technology that difficult and/or improve the
production of large-scale distributed systems;
-
Provide a comprehensive
overview of existing software engineering techniques that may successfully be
applied to deal with the complexity associated with realistic multi-agent
software.
Topics of
Interest.
The workshop is intended to cover wide ranges of topics of software engineering
for large-scale multi-agent systems, from theoretical foundations to empirical
studies. We welcome the submission of papers in all aspects of agent-multi
software engineering, including the following (but are not limited to):
-
Software
engineering theories for large-scale MAS
-
Comparative
studies between multi-agent systems and object-oriented systems
-
Pitfalls and
learned lessons in the construction of large MAS
-
Experiments
and case studies with large-scale MAS development
-
Advanced
separation of concerns in the context of MAS
-
Design
patterns, design principles, and architectural styles
-
Frameworks
and software architectures
-
Reflective
software architectures
-
Coordination
architectures, infrastructures, and tools
-
Domain-specific languages for MAS
-
Requirements
engineering for MAS
-
Software
reliability engineering and MAS
-
Exception
handling and fault-tolerance techniques
-
Mobility and
security issues in large MAS
-
Software
engineering techniques for resource-bounded MAS
-
MAS
development and pervasive computing
-
Verification
and validation techniques for MAS
-
Software
development environments for real-life MAS
-
Modeling of
large MAS
-
UML
application to large-scale MAS
-
Methodologies for agent-oriented analysis and design
-
Testing and
metrics for MAS
Format
and Submissions. The format
of the workshop will combine short presentations and
focused discussion groups. The number of participants will be between 30 and 40, and it will be
restricted to authors of accepted papers and to a few invited guests. The
submission of papers if followed by a review process by the PC,
with the final decision made by the Workshop Organizers. Prospective participants should submit
electronically a position paper (between 5 and
8 pages) in PDF format via
the web site
http://icse2003-submissions.ira.uka.de:8080/selmas-papers/submit/ by February 15, 2003. The format of submitted papers must follow the ICSE/ACM
conference proceedings guidelines,
including no page numbers. We encourage authors
to present novel ideas, critique of existing work, and practical studies and
experiments, which demonstrate how software engineering techniques can assist
the development of large-scale multi-agent systems.
| Publication of the
Proceedings |
Accepted papers
will become part of the Web proceedings of the workshop and be accessible via
the workshop's Web site. An informal
proceedings will be printed and distributed at the workshop. It is the aim of the organizers to publish
revised papers from the workshop in a LNCS volume after the workshop is held.
Jose Alberto R. P. Sardinha - Primary Contact
PUC-Rio - Brazil (sardinha@inf.puc-rio.br)
Alessandro
Garcia
PUC-Rio
– Brazil (afgarcia@inf.puc-rio.br)
Carlos
Lucena
PUC-Rio
– Brazil (lucena@inf.puc-rio.br)
Jaelson
Castro
Federal University of Pernambuco – Brazil
(jbc@cin.ufpe.br)
Alexander Romanovsky
University of Newcastle upon Tyne
– UK (alexander.romanovsky@ncl.ac.uk)
Paulo Alencar
University of Waterloo – Canada(palencar@csg.uwaterloo.ca)
Donald Cowan
University of Waterloo – Canada(dcowan@csg.uwaterloo.ca)
Jose Alberto R. P. Sardinha
PUC-Rio - Brazil (sardinha@inf.puc-rio.br)
|
