提出了一种基于免疫克隆选择算法的基因表达式程序设计混合算法(CS-GEP)。基因表达式程序设计(GEP)是一种新近提出的遗传程序设计方法,已逐渐成为演化计算新的分支。GEP最为重要的优点在于其具有很强的表达能力,而如何充分利用GEP易操作的特点,提高GEP的群体搜索能力是研究较少的一个重要内容。CS-GEP方法借鉴免疫克隆选择原理重新设计了一种克隆选择学习策略替代原GEP算法的遗传算法搜索策略,数值实验结果表明,CS-GEP较GEP具有更好的问题求解能力。
Gene expression programming (GEP), as a new genetic programming method, is proposed as a new branch of evolution computation. However, less interesting is paid to the use of alternative search strategies within GEP. A novel GEP algorithm based on clonal selection principle (CS-GEP) is proposed. CS-GEP simulated the clonal selection process and correspondingly the new clonal operator and mutation operator are defined, and hence CS-GEP is able to obtain the better ability to explore the solution space. The experimental results on symbolic regression problems show the approach more powerful than GEP in accuracy and efficiency.