面向半导体生产线基于MAS模糊协同的成品率预测方法

发布时间：2018-02-25 22:08

  本文关键词： 半导体生产线 成品率预测 多智能体 模糊聚合 支持向量回归　出处：《计算机集成制造系统》2017年04期 　论文类型：期刊论文

【摘要】：针对传统成品率预测模型中需要大量缺陷信息且极少考虑范围预测的情况,借鉴多智能体思想,研究了一种模糊聚合与支持向量回归相融合的方法,对成品率进行预测。在逐步缩减预测范围的同时,多智能体协同调整学习速率等参数,根据确定好的参数构建多个模糊成品率学习模型;利用模糊规则对多个学习模型的预测结果进行聚合,以提高预测准确性;利用支持向量回归将聚合结果去模糊化,得到最终的成品率预测值。仿真实验表明,该方法预测过程较简便,预测范围更精确,具有可行性。
[Abstract]:In view of the fact that a lot of defect information is needed in the traditional product rate prediction model and the range prediction is seldom considered, a method of fusion of fuzzy aggregation and support vector regression is studied based on the idea of multi-agent. At the same time, the multi-agent adjusts the learning rate and other parameters, and constructs several fuzzy yield learning models according to the determined parameters. The prediction results of multiple learning models are aggregated by fuzzy rules to improve the prediction accuracy, and the final yield prediction value is obtained by using support vector regression to deblur the aggregate results, and the simulation results show that, The method is simple, accurate and feasible.
【作者单位】： 北京化工大学信息科学与技术学院;
【基金】：国家自然科学基金资助项目(51375038) 北京市自然科学基金资助项目(4162046) 高等学校博士学科点专项科研基金博导类资助项目(20130010110009)~~
【分类号】：TN305;TP18
，

本文编号：1535329