Research On Proactive And Reactive Scheduling For Mould Manufacturing Workshop Under Uncertain Environments

Moulds and Dies (M&D) serve as an essential constituent for total equipment manufacturing industry, whose manufacturing abilities have been an important symbol to measure industrial production abilities of a country. Although companies continue to introduce and improve the high-level CNC machining equipment, there is still large gap between China and other advanced industrial countries in the manufacture technology of high-level precision mould, which mainly reflected in the machining accuracy is lower, the manufacturing cycle is generally too long, and the order delivery tardiness is serious. Those factors restrict the international competitiveness of our mould companies. The main reason is that mould companies lack an effective and proper production control approach.M&D manufacturing is single piece production order form with resource-oriented so that the manufacturing process is filled with large variety of uncertainties. Hence, it is difficult to make a reasonable and feasible baseline schedule in the mould shop floor. During the schedule executing, frequent impact of unexpected events will make the baseline schedule infeasible. Therefore, it is necessary to take effective proactive-measures, and reasonable rescheduling and control strategy to ensure that the production plan is completed on time.Supported by the National Natural Science Foundation of China (Grant No.50675039,50875051) and the National High-Tech. R&D Program of China (Grant No.2006AA04Z132), our research focused on proactive and reactive scheduling for mould manufacturing workshop under uncertain environments. The modeling methods for uncertainty were analyzed. The following three issues were discussed in detail:the decision-making mechanism of M&D workshop under uncertain environments, the driven mechanism of rescheduling, and the model and algorithm of scheduling. The implementation of proactive-reactive scheduling decision support system of M&D Flexible Flow Shop will help to improve production planning and control, and enhance the market competitiveness of company. The research mainly includes the following aspects.1. Organizational structure of M&D production based on controller of key parts was proposed. The M&D Flexible Flow Shop under uncertain environments was built. The mathematical models for the three major uncertainties were built, and the proactive-reactive scheduling framework was proposed.2. The combination of processing time based on discrete probability model was analyzed. A proactive scheduling model of three stages flexible flow shop was built, which objective was to maximize schedule stability. Furthermore, a width-variable beam search(VBS) algorithm was presented for solving the proactive scheduling model. Finally, the comparative analysis with the fixed width beam search(FBS) was pursued. And, the computational results show that the presented algorithm has a good balance between solution quality and computation time.3. The operation delay was defined and the calculation method of cumulative task delays was presented. The hybride-driven mechanism based on cumulative delays and even-driven was proposed to deal with recessive uncertainty and dominant uncertainty. The simulation model based on Plant Simulation was built to analyze the properties of the hybrid-driven mechanism. The comparative analysis of periodic rescheduling, event-driven rescheduling and hybrid-driven rescheduling were pursued.4. Two-stage reactive scheduling model of mould workshop was proposed based on the reactive scheduling measures. The two-stage reactive scheduling model includs local repair scheduling and reactive scheduling. The local repair scheduling procedure was implemented when the hybrid-driven mechanism was triggered. In local repair stage, the stability was priory consideration. The sequence of job on the machine was keep the same, but the processing time of the affected operation was delayed along the time axis. The constraint propagation tree was defined and affected operation local repair algorithm was proposed. The comparative analysis of the local repair algorithm and Right Shift rescheduling was pursued.5. The new schedule generated by local repair algorithm will be evaluated by evaluation model and the dispatcher. If new schedule caused serious tardiness or high tardiness penalty costs, it needed to go to the second stage of reactive scheduling. The multi-object reactive scheduling model that simultaneously considered efficiency and stability was built. A modified multi-objective genetic algorithm for reactive scheduling model was proposed, which combined the Euclid crowded distance based non-dominated sorting and clustering algorithm based elite archiving strategy.The comparative analysis of modified multi-objective genetic algorithm and NSGA-Ⅱ was carried out. Simulation results show that because of the Euclid crowded distance based non-dominated sorting, MMOGA algorithm can obtained more uniform Pareto distribution than the NSGA-Ⅱ algorithm, can obtained closer approximation of the Pareto optimal front-end;6. Enterprise business processes and production planning hierarchical coordination control mode was analyzed. The system function module and the database were designed. The flexible flow shop proactive and reactive scheduling decision support system was developed for coordinating the whole production workshop activities.