Combinational Disruption Management In Permutation Flowshop With Sequence Dependent Setup Times

Disruption management of production scheduling, as an important branch of operating research, is an applied science of production management to recover manufacturing system at costs as low as possible. It has long been hot and hard topic in the academic research and is also important management problem concerned by manufacturing enterprise. The permutation flow shop scheduling problems have been widespread focused by scholars at home and abroad because of its strong practical background. When the setup time is sequence dependent, its effect on scheduling scheme can't be ignored, which also increase the difficulty of solving the problem. However, the initial scheduling scheme can't be implemented smoothly, even become no longer feasible, because of the uncertain factors in the production environment. So disruption management is required. At present, the great part of the existing work addresses single disruption event independently. But different types of unexpected disruptions may affect the current schedule simultaneously during the real production. Hence for real-life manufacturing scheduling it turns out to be important to handle this situation.In this paper, a combinational disruption management problem in permutation flowshop with sequence dependent setup times is studied. Six types of disruption events, namely machine breakdown, processing time variation, setup times variation, new job arrival, job priority upgrading, and job cancellation, some or all of them are considered simultaneously. We reschedule in order to minimize the negative effect caused by disrupt events, which will result to jobs'sequence deviation from original plan. Further, a bi-objective combinational disruption management model is built up, considering both makespan and sequence deviation. It can be characterized as Fm|prmu,sijk,post-mgt|f1,(?),f2(?) using the well known three-field notation. This problem has been proved to be NP-hard. Then an enhanced memetic algorithm is proposed for optimizing initial population and balancing global search and local search. Finally, we compare the enhanced memetic algorithm with the classical NSGA-II algorithm,when the proposed initial population construction strategy and random generation initial population strategy are adopted respectively. The advantages of proposed initial population construction strategy and enhanced memetic algorithm are shown in response to different combinations of disturbance events and different scales.The research work is of great significance, both in theory and in practice. On the one hand, this paper will give a certain reference for combinational disruption management model construction. Meanwhile, it also provides new ideas for multi-objective initial population construction strategy and global search and local search trade-offs strategy. On the other hand, it can offers effective and diversified re scheduling schemes to manufacturing enterprise in face of the complex and multiple disruption events, so as to guide practice.