DEVELOPMENT OF THE INTELLECTUAL AGENT-ORIENTED SYSTEM FOR DECISION SUPPORT AT ENTERPRISE

• Main Author: G. Chornous
• Format: Article
• Language: deu
• Published: Publishing Center "Kyiv University" 2014-06-01
• Series: Vìsnik. Kiïvsʹkogo Nacìonalʹnogo Unìversitetu ìmenì Tarasa Ševčenka. Ekonomìka
• Subjects: agent-oriented system; decision support; data mining; hybrid approach; model
• Online Access: http://bulletin-econom.univ.kiev.ua/wp-content/plugins/download-attachments/includes/download.php?id=1732

Actual status of management confirms usefulness and necessity for development of scientific modeling tools for decision-making processes based on distributed artificial intelligence. The paper presents opportunities of the agent – oriented approach to support operative and strategic management decisions at the pharmaceutical enterprise. It is argued that the combination of intelligent agents technology and Data Mining (DM) produces a powerful synergistic effect. The basis of the intellectual agent – oriented DSS (AODSS) is proposed to put a hybrid approach to the use of DM. Hybrid intelligent AODSS is represented numerous network of small agents, it provides concurrent operation execution, solutions distribution, knowledge management. Agents can be divided into groups: data agents, monitoring agents, agents for solutions search, modeling agents, impact agents and presentations agents. The result of research is development of AODSS created as a multi-level system wherein the project, process and environment levels are intercommunicated. The combination of intelligent technologies in AODSS allows involve rules, cases, a wide range of DM methods and models. The paper proposes a variant of AODSS implementation within the real enterprise IT-infrastructure based on SAP NetWeaver. The analysis results of the semi-commercial operation of the system assures that it can improve managerial decisions inasmuch as accuracy, consistency, flexibility, speed together form the basis of actual efficient solutions.