The Study And Application On Dynamic Scheduling Of Automobile Mixed-model Assembly Line

With the diversification of social culture and the greatly enriching of automobileproducts, consumer demands are becoming diversified and personalized. The mode ofmulti-varieties and small-batch production has gradually become the main productionmodel in automobile industry. It is of great theoretical and engineering value to studyproduction scheduling problems under the mixed-model conditions. However, due to thedynamic uncertainty of actual production environment, there is a great gap between theclassic scheduling theory and practical application. The study object of this dissertation isdynamic production scheduling of automobile mixed-model assembly line. Two keyelements of the dynamic scheduling strategy were mainly researched on the basis ofscheduling problem modeling and of real-time production data acquisition.Firstly, the gap between the classic scheduling theory and the practical application wererevealed after reviewing the research status of mixed-model assembly line scheduling anddynamic scheduling. The reasons and countermeasures of the situation were then presented.It is pointed out that the emphasis of this paper will be the dynamic scheduling based onreal-time information.Secondly, two basic problems were studied, i.e. scheduling problem modeling andreal-time production data acquisition. To the former, a scheduling model for minimizingthe make-span was established after viewing the optimization objectives and constraints, aHDPSO algorithm was designed and tested. To the later, the requirements and objectiveswere confirmed by analyzing the uncertain factors of automobile mixed-model assemblyline, the equipment selection and application mode of RFID system were then proposed.Then, rolling windows and automobile sets were progressed by discussing the structuralcharacteristics of PBS. The rescheduling period was designed according to the arrival rateof automobiles. Key events were designed according to the disturbance degree of uncertainfactors. Some insertion rescheduling methods were discussed either. Through the test ofdynamic scheduling module, the methods we proposed were good in performance andefficiency.Finally, conclusion was drawn and future research directions were suggested.