GEP是一种新颖的遗传算法,在函数建模的应用中取得良好的结果。给出计算有效基因长度的伪代码,结合GRCM方法阅读基因,快速计算出染色体的适应值。在算法中增加了参数估计模块,用GEP得到较好模型后,用参数估计模块进行参数优化,试验显示这种混合的GEP方法比传统的最小二乘法、神经网络以及遗传程序设计等方法具有更好的性能。
Gene expression programming (GEP), a novelty genetic algorithm, has brought about a considerable increase in performance in the field of automatic modeling of complex functions. A new method was proposed to compute the gene’s effective length, and then quickly compute the gene’s fitness. Further more, parameters optimization module was used. After a good function was got by GEP, parameters optimization module was used to optimize the parameters. The experiment shows that this algorithm is effective.