Overview of my scientific work
My scientific career started in 1984 with my diploma thesis in the area of machine learning in inspectionplanning (Laboratorium für Werkzeugmaschinen und Betriebslehre, RWTH Aachen). Subsequently thedomain was extended to complete process planning and - together with Pfaff in Kaiserslautern - a researchprototype system was developed, which was presented at the Aachener Werkzeugmaschinenkolloquium in1987.
In my PhD thesis I focused on technical diagnosis: Within the Collaborative Research Center ArtificialIntelligence - Knowledge-Based Systems (Sonderforschungsbereich 314 "Künstliche Intelligenz -Wissensbasierte Systeme") I developed an architecture for knowledge-based diagnostic systems(1987-1992). Different diagnostic and learning methods were compared and validated based on real-lifeapplications.
From 1992-1995 I was responsible for the EU project Inreca for the University of Kaiserslautern. Within thisproject a decision support system was developed and validated in different applications, e.g., prognosis ofwind risk factors in Irish forestry (Coillte), credit scoring (Swiss Bank), reuse of object-oriented software(DaimlerChrysler) , service support for a robot manufacturer (Sepro Robotique), and maintenance supportfor a airplane engine manufacturer (CFM International) . The results of the Inreca project flowed in thedevelopment of a commercial tool for case-based/inductive reasoning Kate tools (Kaidara, Paris) andespecially in the new development of the commercial case-based reasoning tool CBR-Works.
I was responsible for the technology transfer project, funded by the state of Rhine-land-Palatinate, thattransfered the Inreca project results (including the Inreca system) into the development of CBR-Works incooperation with tec:inno (now empolis) (1995-1996).
In my habilitation thesis I focused on the evaluation of case-based reasoning systems. A detailed analysisand com-parison of five commercial tools (carried out in 1994, published in 1995) was the starting point fordefining evaluation criteria and for developing a systematic evaluation approach. Both the criteria and theapproach were validated in 1995 based on the Inreca system. The research results achieved within thescope of my habilitation laid the foundation for a methodology for developing case-based reasoningsystems. Advancement of the underlying research ideas showed the utility of case-based reasoning as amethod for supporting organizational learning in software engineering.
Since 1996 I worked on the methodical integration of the experience factory approach for continuous learningfrom experience with the case-based reasoning approach. This resulted in a detailing and operationalizationof the tasks to be carried out in an experience factory. In addition, it resulted in a model for organizationallyintegrating (learning) knowledge-based systems in practice.
The above described approach was used in numerous projects at the Fraunhofer Institute for ExperimentalSoftware Engineering in Kaiserslautern in the scope of my job as group leader/core competencemanager/department head, e.g., within the IESE Experience Factory that was started in 1999 and is incontinuous operation since 2000. These projects were developed based on an architecture for experiencemanagement systems (Interests), which was instantiated based on CBR-Works/orenge (partially togetherwith empolis) or based on self-development. Interests enables intelligent search for different documents,process descriptions, and lessons learned. Demonstrators of instantiations of the Interests architecture areavailable via Internet . Experimental results achieved within the scope of the IESE Experience Factory confirmthat reuse of project experience ("ask your colleague") works more efficient and more effective in combinationwith our technical reuse infrastructure than using the "pure human based approach" only.
Further experimental results are available for participative introduction and evolution of business processes.These results were achieved during the indiGo project, for which I was responsible as overall projectcoordinator. Within the indigo project the Interests architecture was extended with discourse-centereddiscussion fora and text mining for supporting "e-moderators", using the introduction of new and/or revisedbusiness processes as an example.
In the SKe project Interests was integrated with a formal IT security concept for supporting electronic citizenservices, which was validated in applications with the city treasury of Cologne and with T-Com.
Since 2003 I focus on extending the methodical integration of case-based reasoning and experience factorywith software product-line and multiagent systems. My underlying research vision is partially described in our vision paper.