论文部分内容阅读
针对光刻过程非线性、时变和产品质量不易在线测量的特性,提出了一种基于最小二乘支持向量机预测模型和微粒群滚动优化方法的批次控制预测控制器。通过历史批次样本数据构建光刻过程的最小二乘支持向量机预测模型,解决了复杂光刻过程难以建立精确数学模型的难题,提高了预测模型的精度。通过预测误差的反馈校正和微粒群滚动优化算法求解最优控制律,提高了控制精度。性能分析结果表明,与指数加权移动平均方法及非线性模型预测控制方法相比较,批次控制预测控制器控制器减小了不同批次关键尺寸输出的差异,显著降低了关键尺寸输出的均方根误差,有效抑制了过程扰动影响。
Aiming at the characteristics of non-linear, time-varying and difficult-to-measure product quality of lithography process, a batch control predictive controller based on LS-SVM forecasting model and particle swarm optimization method is proposed. The least square support vector machine prediction model of photolithography process is constructed by historical lot sample data to solve the difficult problem of difficult to establish accurate mathematical model in complex lithography process and improve the accuracy of prediction model. The optimal control law is solved by the feedback correction of the prediction error and the particle swarm optimization algorithm to improve the control accuracy. The results of performance analysis show that, compared with exponential weighted moving average method and nonlinear model predictive control method, the controller of batch control predictive controller can reduce the difference of critical dimension output of different batches and significantly reduce the mean square of critical dimension output Root errors effectively suppress the effects of process disturbances.