一种新型的GEP算法及应用研究  

Novel GEP Algorithm and Its Application

作　　者：陈安升[1] 蔡之华[1] 谷琼[1] 张烈超[1] 

机构地区：[1]中国地质大学(武汉)计算机学院,湖北武汉430074

出　　处：《计算机应用研究》2007年第6期98-100,103,共4页Application Research of Computers

基　　金：湖北省自然科学基金资助项目(2003ABA043)

摘　　要：提出一种适合于GEP表达式树构造的新方法,以及相应的新解码方法(GPED)。通过实验对比,GPED可大大缩短演化时间。提出一种新的算法GPEP,将GPEP应用于碎石桩复合地基承载力预测,结果表明GPEP算法在预测精度和演化效率上均超过遗传神经网络、GP等算法。A new way of construction of expression tree for GEP and a corresponding new decoding method called GPED were presented. Through experimental comparison, GPED can enormously reduce the evolutionary time. And a new algorithm GPEP was further proposed and applied to the problem of predicting load-bearing capacity of composite foundation. The result showes that GPEP outperforms the GA_ANN and GP both in prediction accuracy and evolutionary efficiency.

关 键 词：基因表达式编程 遗传程序设计 解码 预测 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]