遗传算法在薄膜特性参量测量中的应用

The Application of Genetic Algorithm in Thin Film Characters Measurement

利用遗传算法,将光学常量作为参量建立全局优化模型.通过遗传变异进行迭代计算从而优化解.它不但避免了接触式测量对样品的破坏,而且相对于其它算法简单清晰,在保证准确度的同时,具有良好的稳定性.实验结果与理想模型误差在1 nm以内,实际样品与台阶仪的测量结果误差控制在20 nm以内.

By using Genetc Algorithm（GA）, global optimization model was constructed, in which optical constans are measured as parameters. Results were optimized by using genetic and aberrance iterative calculations. The damage of samples which is indeed by contact measurement is avoided. This method is simple and have good stabilization with high accuracy. During experiment the one-layer ideal model error is less than 1 nm, and the real model error according to atomic profiler is less than 20 nm.