A new algorithm of automatic programming: GEGEP

  • Authors:

  • Xin Du;Yueqiao Li;Datong Xie;Lishan Kang

  • Affiliations:

  • Department of Computer Science and Technology, China University of Geosciences, Wuhan, China;Department of Computer Science and Technology, China University of Geosciences, Wuhan, China;Department of Computer Science and Technology, China University of Geosciences, Wuhan, China;Department of Computer Science and Technology, China University of Geosciences, Wuhan, China

  • Venue:
  • SEAL'06 Proceedings of the 6th international conference on Simulated Evolution And Learning
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Gene Expression Programming (GEP) has wide searching ability, simple representation, powerful genetic operators and the creation of high levels of complexity. However, it has some shortcomings, such as blind searching and when dealing with complex problems, its genotype under Karva notation does not allow hierarchical composition of the solution, which impairs the efficiency of the algorithm. So a new automatic programming method is proposed: Gene Estimated Gene Expression Programming(GEGEP) which combines the advantages of Estimation of Distribution Algorithm (EDA) and basic GEP. Compared with basic GEP, it mainly has the following characteristics: First, improve the gene expression structure, the head of gene is divided into a head and a body, which can be used to introduce learning mechanism. Second, the homeotic gene which is also composed of a head, a body and a tail is used which can increase its searching ability. Third, the idea of EDA is introduced, which can enhance its learning ability and accelerate convergence rate. The results of experiments show that GEGEP has better fitting and predicted precision, faster convergence speed than basic GEP and traditional GP.