Agent-Based Knowledge Management for Collaborative Engineering

Abstract Collaborative engineering is knowledge work in the sense that engineers and managers must apply their knowledge to solve their problems and proceed with their work. The knowledge that is applied spans from the basic knowledge they learned from schools, application knowledge acquired from the industry, and experiential knowledge obtained from years of working experience. Most of the current research on collaborative engineering support focuses on providing communication and data sharing support for effective coordination. We argue that in order to increase the productivity of current practice of collaborative engineering, we need mechanisms that can not only facilitate information flow, but also provide active knowledge level support for engineers. Our research on KICAD — a Knowledge Infrastructure for Collaborative and Agent-based Design — attempts to develop a network of intelligent agents that capture knowledge from their associated human engineers and provide knowledge level support to them when needed. Among the issues involved in developing such a framework is the issue of knowledge management — how can we model knowledge, how can agents capture, update, manage, and utilize the knowledge for human support? In this paper, we first briefly introduce the KICAD research program and describe the issue of knowledge management in KICAD. After that we present a general knowledge application model (GKAM), the basic conceptual framework of knowledge management in KICAD. An example of applying GKAM in a prototype system will also be discussed.