An Event-Driven Agent-Based Simulation Model for Industrial Processes.

Process manufacturing industries are complex and dynamic systems composed of several processes, subject to many operations and unexpected events that can compromise overall system performance. Therefore, the use of technologies and methods that can transform traditional process industries into smart factories is necessary. In this paper, a smart industrial process based on intelligent software agents is presented with the aim of providing a technological solution to the specific needs of the process industry. An event-driven agent-based simulation model composed of eight reactive agents was designed to simulate and control the operations of a generic industrial process. The agents were modeled using the actor approach and the communication mechanism was based on the publish–subscribe paradigm. The overall system was tested in different scenarios, such as faults, changing operating conditions and off-spec productions. The proposed agent-based simulation model proved to be very efficient in promptly reacting to different dynamic scenarios and in suitably handling different situations. Furthermore, the usability and the practicality of the proposed software tool facilitate its deployment and customization to different production chains, and provide a practical example of the use of multi-agent systems and artificial intelligence in the context of industry 4.0. [ABSTRACT FROM AUTHOR]