Genetic algorithm to speed up modeling turn-around-time

In computational photolithography modeling, there are limitations for the commonly used parameter optimization algorithms. This situation will be more serious when a large number of physical parameters with large search range are to be optimized. Here, we use genetic algorithm (GA) for the parameter optimization. It can largely shorten the model calibration time, and the model accuracy is promised. In addition, an optimized flow called fast GA is developed to further reduce the model calibration time. Through further integrating GA modeling with the metrology, it can even reduce the traditional metrology and modeling time from 1 month to 4 days.