Agent-oriented Constructivist Knowledge Management

Research in Knowledge Management (KM) has evolved substantially in the past 30 years, coming from a centralized view of KM processes to a distributed view, grounded in organizational and cognitive sciences studies that point out the social, distributed, and subjective nature of knowledge. However, KM systems still face considerable resistance, mainly because they generally impose a specific process instead of fitting in the current practices of the organization. This thesis defends a human-centric view on KM, proposing Constructivism as the theoretical framework to guide the development of KM systems and practices. In general, a constructivist perspective on KM focuses on how knowledge emerges, giving great importance to the knowledge holders and their natural practices. Aiming at observing the compliance of the organizational environment to the principles that characterize Constructivist KM, we propose ARKnowD, an agent-oriented methodology to develop KM systems. ARKnowD places strong emphasis in the earlier phases of software development, supporting the analyst on understanding the organizational environment before actually developing a system. Furthermore, ARKnowD consistently conducts to the design of the proposed solution, modeling the system entities, interaction and internal behavior. In this thesis, we demonstrate the applicability of ARKnowD on the analysis of an organizational scenario. This analysis leads to the proposal and development of a socially-aware recommender system named KARe. The core of the system regards a recommendation mechanism, based on an innovative information retrieval technique presented in this thesis. Our work comprises the description, implementation and evaluation of such mechanism.