Two Upper Bound Constraint Theorems and Algorithms for GEP Phenotype
Traditional gene expression programming (GEP) decoding is based on the phenotype expression tree. In the realization of constructing trees and decoding, there will be some problems, including the choice of express storage data structure for discrete expression tree, statement and definition for the height of expression trees, determination of space size for single layer nodes and so on. In this paper, two upper bound theorems and their proofs, which are for the height of phenotype expression trees and numbers of nodes in single layer of GEP gene, are proposed, based on determined head length of the gene. The theorems are established for two principles of contributing trees,"breadth first"and"depth first", solving the problem of determining related attributes of GEP data structure for expression tree, and an GEP algorithm of constructing trees and decoding, using the object array based on object oriented programming language to express the discrete expression tree ,is presented.
NETL
NSTL
万方数据
维普
