摘要
In contrast to the common assumption, machines can be unavailable in most real-life industrial settings for many reasons. In this paper, we study the scheduling flowshops problem under "condition-based" maintenance constraints to minimize the expected makespan using simulation to tackle the randomness of the problem at hand. We propose an effective meta-heuristic algorithm, namely, hybrid simulated annealing-tabu search, to tackle such an NP-hard problem. The proposed method focuses on non-resumability, where the job needs to completely restart. We additionally suppose that machines suffer from degradation due to shocks; thus, preventive maintenance must be performed on machines. Furthermore, it is assumed that the degradation value of a machine is known at inspection time. Overall, seven adaptations of existing meta-heuristic and heuristic methods are evaluated for the integration of preventive maintenance and are applied to a set of 960 instances. The preliminary results showed that the proposed algorithm of this paper performs better than other existing methods.
- 出版日期2010-1