Research On Joint Scheduling Method Of Production And Maintenance

With the advancement of science and technology,the scale of semiconductor manufacturing has expanded exponentially,and the complexity has gradually increased.Besides,the market competition has become increasingly fierce.How to develop a reasonable production plan,ensure the reliability of equipment and operational efficiency to obtain maximum economic benefits in the fierce market competition is an urgent problem for semiconductor manufacturing enterprises.Production scheduling and equipment maintenance are the two core parts of enterprise operation management.Production scheduling and equipment maintenance are directly related to the cost,quality and delivery time.At present,there is not much research on the joint scheduling of production scheduling and equipment maintenance as a whole.And most of it is to regard equipment maintenance as one of the restrictions of production scheduling.There is less research to quantify the interactions between the two and the effects of variability.The semiconductor parallel production line is taken as the research object,and the defects of the existing production and maintenance joint scheduling method are partly solved,such as inaccurate measurement of production and maintenance,and insufficient consideration of uncertain factors.This thesis focuses on the following aspects:Firstly,production scheduling and equipment maintenance are taken as a whole,and the impact and correlation between production scheduling and preventive maintenance is quantified.A joint maintenance strategy based on failure rate-based preventive maintenance,minor repairs for minor failures during processing,and direct replacement of faults before the start of machining is combined with production scheduling.The impact of production on maintenance is measured through the deteriorating factor in the process and the impact of maintenance on production is measured through the pm improvement factor.Based on this premise,a joint scheduling model with the goal of minimizing makespan is established.Secondly,common algorithms to joint scheduling model has been analyzed and the genetic algorithm is selected to solve the problem according to the mathematical characteristics of the joint scheduling model and the characteristics of the uncertainty and complexity of the research object.A constrained coding method is designed.Operators and parameters are redesigned to solve the problems of “premature” and low search efficiency.The elitist preservation strategy is used to maintain the best solution found over time before selection.Then the validity of the model and the designed algorithm of this thesis is verified by several examples.Finally,the joint scheduling prototype system of semiconductor parallel production line is designed and implemented based on the joint scheduling model and algorithm.Through the system,information management can be realized,and a joint scheduling sequence of production and maintenance can be realized.Therefore the efficiency is improved and the cost is saved.