FUZZY NEURAL NETWORKS BASED RESCHEDULING STRATEGIES OPTIMIZATION OF SEMICONDUCTOR FABRICATION LINE

Zhang Jie  Zhai Wenbin  Yan Juanqi  Ma Dengzhe

(Institute of Computer Integrated Manufacturing, Shanghai Jiaotong University, Shanghai 200030)

Abstract: For the lack of rescheduling optimization approach for semiconductor fabrication line status in quo, a novel fuzzy neural networks based algorithm for rescheduling strategies optimization of semiconductor fabrication line is proposed. The rescheduling strategies are divided into line layer, machine group layer and machine layer. The samples for FNN are setup by simulation evaluation. The relation between the input of FNN, such as disturbance, system state parameters, and output of FNN, optimal rescheduling strategy, is built by FNN. With an illustrative example of a 152 mm semiconductor fabrication line in shanghai, the lineal regression between FNN output and targets demonstrates the effectiveness of the approach.

Key words: Semiconductor fabrication line  Rescheduling  Fuzzy neural networks

CLC No: TH166

国家自然科学基金资助项目(50375098). Received 20041112, received in revised form 20050329