“Our studies discover that multi-agent systems (MAS) are an effective means for offering novel solutions for BI-systems to manage the large volume of distributed data, the latency of data extraction and the delay in the analysis processes of data,” the author explained. Integrated researches between agent-technology, BI, and distributed data mining (DDM) have become popular in last decades. Due to complexity of such systems, researchers facing challenges during the development life-cycle. “In the context of agent-based system engineering, a number of agent-oriented methodologies exist. Each of these methodologies comprises several engineering phases to guide developers to produce various models during the development process for agent-based systems,” she added. Conversely, existing agent-oriented methodologies cause challenges for developers during the selection processes due to the diverse modeling notations, processes, specifications, and principles. Therefore, none of these agent-oriented methodologies have been accepted as a standard for developing such agent-based complex systems.
“In the content of new designing approach, we propose three important steps that are not discussed, or considered in literature before,” she counted. These are: (1) Evaluation processes of agent-oriented methodologies to find suitable methods for the development of agent-based BI-systems; (2) an implementation of existing agent-oriented methodologies during the analysis- and design-phases for agent-based BI-systems shortly called BI-MAS, and finally, (3) a demonstration of a new approach for verification and validation (V&V) processes to test proposed model for BI-MAS. Karima added that the research is based on the engineering of data management of BI together with artificial-intelligence (AI) society modeling. This novel design approach provides more motivation for researchers in these areas to offer very new and technical solutions for resolving existing challenges of traditional BI-systems.
The public defence of the doctoral thesis “A New Design Approach for Multi-Agent Based Business Intelligence Generation” took place on July 4. The dissertation was supervised by Associate Professor Alexander Horst Norta (Tallinn University of Technology) and Prof. Tobias Ley (Tallinn University). The opponents are Prof. Josef Küng (Johannes Kepler University Linz) and Prof. Kuldar Taveter (Tallinn University of Technology). The defence is held in English.
The full thesis can be read at the TU Academic Library e-depository ETERA.