Data collection for global monitoring and trend analysis in the GRACE multi-agent system

Multi-agent systems (MAS) paradigm is a suitable approach to implement distributed manufacturing systems addressing the emergent requirements of flexibility, robustness and responsiveness. In such systems, the distributed agents have a local view of the system environment being the global data collection a complex and critical issue to provide the functionalities of the ISA-95 standard, such as dynamic scheduling, maintenance management and quality assurance. This paper describes the data collection mechanisms in a multi-agent system that supports the global monitoring over the time and the trend analysis to perform self-adaptation and self-optimization. The proposed approach was developed and tested in an industrial washing machines production line.