Enhancing Intelligent Knowledge Systems Using Organization-Centered Agent Models

Abstract

Many construction industry–specific, agent-based applications have been deployed to enhance knowledge systems. These applications are generally agent-centered multi-agent systems (ACMAS). Interaction between ACMAS agents is designed around the internal mental states of an agent and the relationship between these states and its overall behavior. Communication in this approach comprises speech acts whose meaning can be described in terms of the mental states of an agent. This interaction protocol makes it difficult for agents that have been implemented using different approaches to communicate. This paper reviews the potential for addressing this limitation through using organization-centered multi-agent systems (OCMAS) in which agent interaction is designed around macro-level concepts such as organizations, groups, communities, and roles. Although the use of organizational metaphors in agent-based application is still in its infancy, there are some emerging modeling tools that can be used to advance the existing agent-based applications into open MASs that exhibit the dynamic and flexible characteristics of open, distributed systems. The paper illustrates this potential by using a proof of concept based on decision support requirements for highway maintenance management.