Research On Scheduling Optimization Of Photolithography Workshop With Resource Constraints

Scheduling problem is one of the most important decision-making processes in manufacturing system.In the semiconductor manufacturing industry,the high equipment investment of lithography machine makes it become the bottleneck of the system.Different from the traditional production scheduling problem,mask is needed as an auxiliary resource in lithography.Due to its scarcity,the mask can be transferred from one machine to another through the material transfer system.The problem involves jobs,machines,and masks,and is constrained by sequence related adjustment times and position related mask transfer times.This way of sharing resources provides a more flexible production mode for manufacturing enterprises,but it also makes the scheduling problem more complex.This paper starting from the existing production scheduling methods,introduces the advantages and disadvantages of the existing solutions in detail,and the characteristics of wafer processing scheduling problem in photolithography workshop,analyzes the impact of the time cost of mask plate in the process of processing scheduling due to its increase in replacement and transfer process,and at the same time,considering the requirements of the wafer for its special photolithography machine and mask plate,constructs the corresponding scheduling The model is solved by the improved mole algorithm.The detailed research contents of this paper are as follows:(1)In this paper,the scheduling problems and scheduling methods of photolithography workshop at home and abroad in recent years are sorted out,which provides a reference for the analysis and solution of the following problems.(2)This paper studies the single machine scheduling problem of photolithography workshop with mask adjustment time constraints,analyzes its characteristics,establishes an integer programming model with the objective of minimizing the total weighted completion time,and designs an improved genetic algorithm to solve the problem,which lays a foundation for the subsequent establishment of the model with complex constraints.(3)Aiming at the scheduling problem of uncorrelated parallel machines in lithography workshop with two resource constraints,two modeling methods based on network variables and linear sorting are considered.Aiming at minimizing the total weighted completion time,a mixed integer programming model is established,and the advantages and disadvantages of the two modeling methods are compared from the theoretical and experimental perspectives.The results show that for this problem,the integer programming with the separation of allocation and sorting variables has higher computational efficiency than the mathematical model with network variables.(4)By using the mole framework,aiming at the scheduling problem of unrelated parallel machines in the lithography workshop with two resource constraints,the process decoding rules based on the greedy rules are designed according to the characteristics of the problem,and the original mole algorithm is optimized and improved to solve the problem by combining the genetic algorithm and the operators of variable domain search.The correctness of the algorithm is verified by comparing the results of the proposed improved mole algorithm with the small calculation examples of the gurobi model of the mathematical model.In addition,compared with the results of variable neighborhood search algorithm and genetic algorithm,the algorithm shows good performance.