Agent-based, hybrid control architecture for optimized and flexible production scheduling and control in remanufacturing

Introduction:
Motivated by high ecological and economical potentials and driven by new laws, remanufacturing is receiving increasing attention as a process that puts used products into “as good as new or better” condition. Within this process, there are many challenges, which are unseen in manufacturing, such as the uncertainties resulting from unknown conditions of the used products. This places special demands on the control of the remanufacturing system (RS).
Methodology:
To handle these uncertainties an agent-based hybrid control architecture comprising centralized and decentralized components is presented. In the former, the scheduling takes place including the consideration of the use of automated guided vehicles (AGV) to realize flexible material handling within the RS. The scheduling of machines and AGVs is thereby considered simultaneously and not separately, as it is the case in currently available control systems. For the optimization of the simultaneous scheduling Constraint Programming (CP) is used. In the decentralized component, all participants within the RS will be networked as a cyber-physical system and controlled by respective agents. These agents can communicate with each other in order to find solutions. The architecture is implemented as a multi-agent system.
Results:
Simulation results, using benchmark instances, show that simultaneous scheduling results in a 19.7% reduction of the makespan. Furthermore, the CP-based approach delivers the best results, compared to other approaches for simultaneous scheduling, which are also achieved in a significantly shorter computing time.