Research Project CoSa

Component based framework for an effective and efficient simulation of (agent) systems


     Runtime:

     Project coordination:

     Scientific staff:





     Student staff:







    

     Source of funding:

Jan 01 2003 until Dec 31 2013

Prof. Dr. Adelinde M. Uhrmacher

Dipl.-Inf. Alexander Steiniger • M.Sc. Tobias Helms
Dr.-Ing. Jan Himmelspach • Dr.-Ing. Roland  Ewald • Dr.-Ing. Stefan Leye • Dr.-Ing. Stefan Rybacki • Dipl.Inf. Mario Krahmer; • Dipl. Inf. Susanne Biermann • Dr.-Ing. Mathias Röhl • Dr.-Ing. Matthias Jeschke • Dr.-Ing. Mathias John

Johannes Rössel • Valerius Weigandt •Mathias Rimane • Enrico Seib • Dortje Löper • Ulricke Borchardt • Robert Waltemath • Simon Bartels • Johannes Becherer • Thomas Beer • Gabriel Blum • Nico Eggert • Roland Ewald • Stefan Friedrichs • Florian Gewandt • Lydia Jost • Sven Kluge • Stefan Leye • Sebastian Lieske • Frank Manteufel • Steffen Maas • Rene Michalski • Thomas Nösinger • Christian Ober • Björn Paul • Jan Pommerenke • Kathrin Rohloff • Oliver Röwer • Hans Schipke • Rene Schulz • Carl Tuemmler • Felix Willud • Felix Woitzel • Martina Gierke

DFG  (DFG GEPRIS Link)

Abstract

The goal of this project is to develop a modeling and simulation framework which is sufficiently flexible and modular to support research on modeling and simulation methods, the design of domain dependent simulation tools, applications in diverse areas, and modeling and simulation of complex, heterogenous (agent-based) systems.

The goal of the project "CoSA", funded by the DFG, is the creation of a component-based simulation system for agents. The original project idea was based on the insight that agent systems are software systems for which simulation plays an increasing role in investigating their functionality and performance. Thereby, the properties that are typical for agents, such as flexibility, autonomy, resource intensity, and mobility, require the development of tailored modeling and simulation methods. Based on previous work in agent-based simulation and current developments in modeling and simulation, a simulation system comprising reusable components shall be developed, which will enable efficient and effective experimentation with multi-agent systems in virtual, dynamical environments. To conduct such simulation studies, flexible simulation systems are required. Flexibility refers to possibilities in designing, controlling and supporting the interpretation of experiments. The efficiency of such simulation studies not only depends on efficient, possibly parallel-distributed execution of test runs, but also on the ease of developing new test scenarios. This is the area of tension between efficiency and effectiveness that shall be researched with the component-based framework (JAMES II) to develop as well as selected, concrete simulation studies with multi-agent systems. In developing this framework, the strong focus on multi-agent systems was increasingly dropped. Models for multi-agent systems are "only" a special application from the point of view of a general, open, and reusable modeling and simulation framework. Besides specialized model description languages, there seems to be no necessity for specialized software products for modeling and simulation of multi-agent systems - as long as the general software products can offer the existing characteristics of models of multi-agent systems. For the developed, open framework, extensions were developed that allow for the effective and efficient modeling and simulation of multi-agent systems. Further research results achieved in this project support this, but are equally available for every other modeling paradigm. Thus, the solution developed surpasses the initial idea, but still completely contains it.