Multi-agent architecture and framework that eases the development and the operation of large-scale, distributed applications and services.

Lightweight multi-agent architecture and framework that simplifies the development of applications for different kind of devices.


Intuitive and extensible monitor for controlling distributed multi-agent infrastructures at runtime.

Well-equipped BPMN editor, powerful multi-language transformation and workflow simulator.

ASGARD 2.4 Released

We are very happy to announce the release of ASGARD 2.4. The new version is built for the recently released version 5.1.5 of JIAC. Main new feature is the integration of searching abilities that allow to search for property values of any entity displayed by ASGARD. Also, the visualization of different types of communication by color schemes from the last beta is now part of the release.

A complete list of changes can be found in the changelog contained in the release package.

Download the new version from the ASGARD project page and enjoy watching your JIAC agents at work!

Note to mac users: If you are using Java 7, you might experience that after maximizing the ASGARD main window, mouse over texts will not appear over entities and behavior on mouse clicks is erratic. This is caused by a bug in the OSX version of the Java 7 JVM. Normal behavior can be restored by clicking the window title or switching to another application window and back. We hope that this bug will be fixed in future releases of Java 7.

JIAC 5.1.5 Released

We are very happy to announce the release of version 5.1.5 of the JIAC multi-agent framework. In addition to the usual bugfixes, stability improvements and dependency updates, JIAC 5.1.5 improves interoperability of JIAC agents by enhanced directory data exchange features.

Note: In JIAC 5.1.5, the AgentBean property executeInterval was renamed to executionInterval to be consisted with the property name of  Agents in the configuration. The old name can still be used in this version, but will trigger a warning message in the log output. Please change the name of the property in the configuration files of your JIAC applications to get rid of the warning and be prepared for future JIAC releases.

As usual, JIAC is available from our dai-open maven repositories.

Enjoy working with agents!

The JIAC development team

ASGARD 2.4-beta Available

We are proud to announce that the beta version of ASGARD 2.4 is now available for download. ASGARD 2.4-beta’s main new feature is the use of different colors for visualizing communication, which allows to differentiate between simple message exchange and action invocations. Furthermore, ASGARD 2.4-beta supports the full JIAC 5.1.4 API. For a full list of changes, please consider the change log provided in the archive. It is available as standalone package for Windows, Linux and Max OS X and supports both Java 6 and Java 7.

[Download ASGARD 2.4-beta]

[ASGARD description page]

JIAC in Action: PIA

PIA, the Personal Information Assistant currently under development by DAI-Labor, is powered by the JIAC platform and makes excessive use of JIAC’s distribution and communication features. What this means can be seen in this image, which shows what is happening between JIAC agents and Agent Nodes when PIA is at work.


ASGARD Screenshot of the running PIA implementation


PIA is an agent based enterprise and desktop search system that allows every employee of a company to maintain and search information on their desktop computer and their company’s network. The system analyzes and frequently updates its knowledge base and acts as a platform to satisfy the users’ information need. If a user explicitly states their interests in a profile, the system informs the user about changes and updates that are relevant to these interests.

[Try PIA here]

JIAC 5.1.4 Released

The JIAC Developer team is proud to announce the release of JIAC 5.1.4. The new version features various stability improvements, dependency updates and bug fixes. Furthermore, new features have been added to support building large scale applications.

As usual, JIAC 5.1.4 is available for maven integration from our dai-open maven repositories.

Enjoy working with agents!

The JIAC Developer team

Bugfix Release: ASGARD 2.3.1

As the recent update of Java 6 for Mac OS X (Java 1.6.0 Update 51 / Java for OSX 2013-04) turned out to be incompatible with ASGARD 2.3. we have decided to release a fixed version of ASGARD for all users running Java 6 on Max OS X. Mac users that have already migrated to Java 7, as well as users of all other operating systems, seem not to be affected, so there is no need for them to update to ASGARD 2.3.1. Symptoms of the problem in Mac OS is that ASGARD shows a completely empty window and optionally a dialog window indicating a graphic device initialization error. Note that these problems only occur when Java 1.6.0 Update 51 is used, all prior versions are not affected by this problem.

ASGARD 2.3.1 can be downloaded [here].

ASGARD 2.3 Released

We are glad to announce the release of ASGARD version 2.3, built to support JIAC 5.1.3. This release of ASGARD is finally able to support Java 7 on OSX by using LWJGL 2.9.0. All other additions, changes, and bugfixes in the beta versions since 2.2 are included in this release.

Get the new version [here].

Changes since last stable release (2.2):

New Features:

  • Added feature to deploy agents at runtime.

Fixed Bugs:

  • Fixed network interface binding on windows.
  • Fixed jconsole location in linux environments.
  • “-useLocalhostUrls” parameter was not parsed properly.
  • Correct height for communication channel objects when the framework visualization is resized.
  • Fixed issue of messages appearing at wrong positions.
  • Limiting maximum number of message visuals to avoid memory flooding.
  • Using Thread pool for connection checks to avoid flooding the system with threads.
  • Fixed blocked RMI Registry connection dialog.
  • Correct Zbuffer bit depth for Intel HD Series graphic adapters.
  • ASGARD is not blocked anymore during RMI registry connection attempts.


  • Framework captions now use the DAI Corporate design font, which is Titillium.
  • Using only dependencies from public repositories to make source releases actually buildable for users.
  • Creating own multicast receivers for each network interface instead of relying on address
  • Updated to LWJGL version 2.9.0 to provide Java 7 compatibility in OSX.
  • Moved new connection management to own class and added connection attempt counter display.
  • If no port is specified when connecting a registry, the default port (1099) is used.
  • Localhost URLs are filtered per default before attempting to connect.
  • Failed connection attempts are not visualized as gray agent nodes per default.
  • Using weak references in channel visuals to avoid “zombie” objects or visuals.
  • Changed lwjgl version to 2.8.3, new native libraries.
  • Optional kiosk mode for presentations on public screens.

Distributed Evolutionary Optimisation for Electricity Price Responsive Manufacturing using Multi-Agent System Technology

Our paper “Distributed Evolutionary Optimisation for Electricity Price Responsive Manufacturing using Multi-Agent System Technology” will be included in the next issue of the International Journal On Advances in Intelligent Systems. In this paper, a distributed optimization system using the JIAC agent framework is presented.

Authors: Tobias Küster, Marco Lützenberger, Daniel Freund, Sahin Albayrak

Abstract: With the recent uptake in renewable energies, such as wind and solar, often comes the apprehension of unreliable energy supply due to variations in the availability of those energy sources, also resulting in severe fluctuations in the price of electricity at energy exchange spot markets. However, those fluctuations in energy costs can also be used to stimulate industry players to shift energy intense processes to times when renewable energies are abundant, not only saving money but at the same time also stabilising the power grid. In previous work, we presented a software framework that can be used to simulate and optimise industrial production processes with respect to energy price forecasts, using a highly generic meta-model and making use of evolutionary algorithms for finding the best process plan, and multi-agent technology for distributing and parallelising the optimisation. In this paper, we want to wrap up our work and to aggregate the results and insights drawn from the EnEffCo project, in which the system has been developed.

Source: International Journal On Advances in Intelligent Systems [link]

Semi-Automated Generation of Semantic Service Descriptions

Our paper “Semi-Automated Generation of Semantic Service Descriptions” has been presented at the “Self-Explaining Agents” – Special Session during the PAAMS 2013 conference at Salamanca, Spain. The paper describes an approach of semi-automated enhancement of multi-agent system actions with semantic service descriptions providing a first step towards automated service composition.

Authors: Nils Masuch, Philipp Brock, Sahin Albayrak

Abstract: The increasing complexity and dynamics of distributed systems make the management and integration of new services more and more difficult. Automation processes for the definition, selection and composition of services for goal achievement can produce reliefs. However, a high degree of self-explanation of the services is obligatory for this. Today’s multi-agent frameworks only provide insufficient solutions to this. Within this paper we will outline an approach, which enables the integration of semantic service descriptions into multi-agent systems with reasonable effort.

Source: PAAMS 2013 Homepage

Engineering Industrial Multi-Agent Systems: The JIAC V Approach

Our paper “Engineering Industrial Multi-Agent Systems: The JIAC V Approach” will be presented at the Engineering Multi-Agent Systems (EMAS) Workshop during the AAMAS 2013 conference at St. Paul, MN, USA. This paper describes JIAC V and its features aiming towards industrial requirements in greater depth than our paper “JIAC V – A MAS Framework for Industrial Applications”.

Authors: Marco Lützenberger, Tobias Küster, Thomas Konnerth, Alexander Thiele, Nils Masuch, Axel Heßler, Jan Keiser, Michael Burkhardt, Silvan Kaiser, Jakob Tonn, Sahin Albayrak

Abstract: The community of agent researchers and engineers has produced a number of interesting and mature results. However, agent technology is still not widely adopted by industrial software developers or software companies. Yet, given that the software paradigms which are currently employed by the software industry, such as service-oriented architectures or cloud computing, have much in common with agent- oriented software engineering, industrial software projects could greatly benefit from agent technology. In this paper, we make an approach to analyse the requirements of current industry-driven software projects and show how we were able to cope with these requirements in the JIAC V agent framework. We argue that the lack of industry-grade requirements and features in other agent frameworks is one of the reasons for the slow acceptance of agent technology in the software industry. The JIAC V framework tries to bridge that gap – not as a final solution, but as a stepping stone towards industrial acceptance.

Source: Engineering Multi-Agent Systems (EMAS), AAMAS 2013 Workshop, Saint Paul, USA.