Securing Agent-Oriented Systems: An Argumentation and Reputation-based Approach

In this paper, we propose a new model for securing agent-based systems in which agents are equipped with reasoning capabilities allowing them to interact with each other. The agents can reason about the reputation of each other using their argumentation systems. The reputation is dealt with as a quantitative value computed using a set of parameters based on the interaction histories and the notion of social networks. The problem of securing autonomous interacting agents in a distributed setting is core to a number of applications, particularly the emerging semantic grid computing-based applications such as e-business. Current approaches fail to adequately address the challenges of security in these emerging applications. These approaches are either centralized on mechanisms such as digital certificates, and thus are particularly vulnerable to attacks, or are not suitable for argumentation-based agent systems in which agents use advanced reasoning capabilities.

• Type: Conference
• Year: 2007
• Where: ITNG
• Authors: Jamal Bentahar, John-Jules Ch. Meyer, Bernard Moulin
• Keywords: Agent-based Systems, Autonomous Interacting Agents, ITNG 2007, Reasoning Capabilities