摘要： 介绍了基因表达式程序设计方法的基本原理，针对求解复杂函数模型反问题中经典GEP算法多样性表现不足，甚至出现早熟的问题，提出了一种基于动态变异算子的改进的GEP算法——IGEP算法，从理论上对该改进算法进行了复杂度分析和收敛性分析。通过求解复杂函数模型反问题的多个实验将改进算法与传统方法、神经网络方法、经典GEP算法进行了对比，结果表明：该方法建立的复杂函数反问题拟合模型比经典GEP方法、传统方法、神经网络方法得到的模型更加优秀。

关键词: 基因表达式程序设计, 自动建模, 复杂度, 收敛性, 参数模型

Abstract: The basic principle of gene expression programming (GEP) is introduced. An improved GEP algorithm called IGEP based on dynamic mutation operator which deals with the problem of complex function auto modeling of complex function is presented, the algorithm complexity of the IGEP is given. Many simulation results show that the models set up by the paper are better than the models set up by classic GEP, traditional algorithm and nerve network.

Key words: Gene expression programming(GEP), Auto modeling, Complexity, Convergence, Parameters model

中图分类号: 

• TP183