摘要： 在基因表达式编程的基础上提出2种新的解码方法，分析了它们的时间和空间复杂度。第1种方法完全遵照原始基因表达式编程中基因型与表现型之间的映射关系，直接在基因型上计算和求解表达式。第2种方法在基因结构保持不变的前提下，利用栈来计算和获取表达式，使得基因型与表现型之间的映射关系已不同于前者。这2种方法对重组算子有着不同程度的影响。在3组数据上的实验结果表明，2种新的解码方法可行而且高效，第2种解码方法更有利于优良子树模式的保护。

关键词: 基因表达式编程, 解码, 基因型, 表现型, 复杂度

Abstract: Two new decoding methods for Gene Expression Programming(GEP) are proposed in this paper, and their time and space complexity are analyzed. The first method conforms to the mapping relationship between genotype and phenotype of GEP, and the expression and its value are obtained directly on genotype. The second method gets the expression and its value by stack, which makes the mapping relationship between genotype and phenotype different from the foregoing. The two methods have different effects on the recombination operators. The experiments on three groups of data show the new decoding methods are feasible and efficient. The second method is more propitious to preserve the excellent schemes.

Key words: Gene Expression Programming(GEP), decoding, genotype, phenotype, complexity

中图分类号: 

• TP311.5