Towards an Agent-Based Knowledge Management System - Developing Concepts in Applied Intelligence - page 95

With agents, it is possible to capture the representation, coordination, and cooperation
between heterogeneous processes and other agents. An appropriate KMS should have
the ability to extract proper tacit knowledge when requested to do so [9, 10, 11] and
has to support the process of DM and Organizational Learning (OL).
In this paper, we propose a KMS that employs agents and Case-Based Reasoning
(CBR) to extract tacit knowledge and support OL and decision making at the level of
individual worker, group workers and the whole organization. Agents will perform
the tasks of inquiry, investigation, sharing and updating of knowledge bases, includ-
ing previous similar cases, in order to help in finding the required knowledge for a
specific problem. We have applied the system to develop a banking training assistant
system and have run several experiments to measure a number of functionalities such
as reliability, efficiency, the impact of weight parameters on extracted results, and the
degree of match between the results produced by our system and the opinions of ran-
domly selected experts in the chosen application. Section 2 will be concerned with
Multi-Agent Systems (MAS), OL and CBR. In section 3 we present an agent-based
learning KMS. Section 4 will be concerned with data collection and implementation
of a training assistant system for the central bank of Jordan.
2 MAS, OL and CBR
Agents are specialized problem solving entities with well-defined boundaries with the
ability to communicate with other agents. They are designed to fulfill a specific pur-
pose and exhibit flexible and pro-active behavior. They are autonomous in the sense
that they can operate on their own without the need for guidance. They have control
both over their internal state and over their actions.
A MAS can be defined as a collection of agents with their own problem solving
capabilities and which are able to interact among them in order to reach an overall
goal [5]. KM is mainly concerned with using, spreading, sharing, representing and
storing the available knowledge of an organization. OM is about how to collect, store,
and provide access to experiences and skills of both stored records and tacit knowl-
edge. OL is concerned with enhancing the organization problem solving process [13].
In [3], a model of learning is proposed. It includes three learning levels: individual,
group and organizational and two routes of flow: upward (individual to organiza-
tional) and downward. This suggests that knowledge be structured to capture the three
learning levels.
CBR is a mechanism for solving new problems based on the solutions of similar
past problems [7]. CBR supports many of the activities of KM and has been fre-
quently proposed as a methodology for KM applications. CBR has been employed in
the design of many KM systems [ 2 ].
3 Towards an Agent-Based Learning KMS
An important feature of a KMS is Knowledge Communication (KC). KC allows for
making knowledge accessible to and reusable by its different components [10, 11,
12]. We employ four types of intelligent agents: (1) Knowledge Cooperative Agent
(KCA) that will be responsible for analyzing and searching the Knowledge-Bases