Maintenance operations have a direct influence on production performance in manufacturing systems. Maintenance task prioritization is crucial and important, especially when availability of maintenance resources is limited. The decision on task assignment is often made through heuristic methods or experience, which could cause more downtime and the production losses. In this paper, a new maintenance task prioritization policy based on data driven bottleneck detection and reliability-based maintenance opportunity window calculation is introduced. An experiment in simulation of a real production line shows the proposed policy is able to improve the system reliability, increase the throughput and minimize the total cost of system operation.
An Enhanced Data-Driven Algorithm for Shifting Bottleneck Detection