An intelligent immune agent management system for facilitating e-diagnosis in manufacturing

This paper focuses on designing and implementing an agent framework for the purposes of electronic prognostics of critical tools and machinery in the plant. It deals with dual communication i.e. one originating from the sensors and relayed to various reception centers and the other originating from an intelligent prognostics engine (IPE) being fed back to the control mechanism (manual/automated) of the machinery. The paper presents the feasibility of deploying agents in embedded and supervisory controllers for collating, managing and communicating information. The conceptual design and prototype model establishes the efficacy of the agent framework in diagnostics as well as prognostics. The framework is developed using Java agent development framework (JADE). This paper concentrates on development of peer agent communication, communication ontologies; agent behaviors based on messaging performatives, integration issues and graphical user interfaces for diverse agents and human interfaces. It also evolves a conceptual agent framework for manufacturing environment. The work can potentially improve reliability: reduce equipment downtime by using intelligence and contextual information from sensors and events. It facilitates the timely availability of the critical data for real-time monitoring and corrective signals, as well as a transparent and rapid configuration mechanism for intelligent monitoring.