Eindhoven University of Technology
Assembly Line
A simulation-based approach to design an automated high-mix low-volume manufacturing system
In this paper, we address the profit optimization problem of an automated high-mix low-volume manufacturing system, which originates from a real-world problem at our industry partner. The manufacturing system includes buffer units from which jobs are automatically transported to workstations, i.e., using automated material handling devices. We consider three different automation concepts for the system: (1) a configuration with parallel buffers and a dedicated robot to work them, (2) a configuration that employs shared buffers that are tended to by automated guided vehicles (AGVs), and (3) a proposed hybrid configuration that takes elements of both aforementioned configurations. We propose a simulation-based approach, which uses simulated-annealing (SA), enriched with the reduced variable neighborhood search (RVNS), to determine the best system configuration for a high-mix, low-volume manufacturer. Decisions concern the choice of automation equipment and the capacity of both parallel and shared buffers. We illustrate the efficacy of the proposed hybrid concept and the proposed SA-RVNS approach with an industry case study using real-world data from our industry partner. Our analysis shows that the proposed concept increases the profit by around 10–30% compared to the others, and the AGV travel time plays an important factor in the proposed concept to yield its true potential.
Motion Control, Mechatronics Design, and Moore's Law
Technology in a broad sense is driven by developments in semiconductor technology, particularly with respect to the computational power of devices and systems, as well as sensor technology. The progress of semiconductor technology has demonstrated an exponential curve since the middle of the previous century, representing Moore’s Law. Consequently, it is of utmost importance to bridge the gaps between disciplines in the fields of control, automation, and robotics. Moreover, data-driven approaches need to be combined with model-based design. This will lead to new digital twinning and automated design approaches that provide major opportunities. Furthermore, this necessitates the redefinition of our university system.