Industrial risks management using agents and dynamic case-based reasoning

This article presents the architecture of a multiagent system using dynamic Case-Based Reasoning for the evaluation of potentially risky situations in an industrial park. This architecture allows a real-time comparison of the observed situation with past situations describing crises stored in a case base. It implements a dynamic multiagent Case-Based Reasoning. Having presented the goals of the system, we make a comparison of the proposed approach with CBR techniques for dynamic situations. We present then in detail the behaviour of the agents used in the architecture.