Agent-based modeling and simulation of an autonomic manufacturing execution system

Production management systems must constantly deal with unplanned disruptive events and disturbances such as arrivals of rush orders, raw material shortage/delays or equipment breakdowns along with a multitude of interactions in the supply chain which constantly demand on-line task rescheduling and order execution control. For responsiveness and agility at the shop-floor, a distributed design for manufacturing execution systems is proposed based on autonomic units that fill the gap between production planning and shop-floor control. An interaction mechanism designed around the concept of order and resource agents implementing the monitor-analyze-plan-execution loop is described. Generative simulation modeling of an autonomic manufacturing execution system (@MES) is proposed in order to evaluate emerging behaviors and macroscopic dynamics in a multiproduct batch plant. Results obtained for an industrial case study using a simulation model of the proposed @MES are presented. The usefulness of agent-based modeling and simulation as a tool for distributed MESs design and to verify performance, stability and disturbance rejection capability of an interaction mechanism is highlighted.

► Detailed design of an interaction mechanism for distributed production control.
► Generative simulation for behavior verification in multi-agent systems for MES design.
► Introducing the autonomic computing paradigm in shop-floor automation.
► Simulation results highlight robustness and stability of the autonomic MES.