Research And Applications Of GEP In Symbolic Regression

Gene Expression Programming is a sort of burgeoning search algorithms: It has remarkable difference with traditional GA and GP such in form of result, expression and processing of individual and so on. Based on the foundation of GA, GP develops fast. Though it surpasses the genetic algorithm in performance, it gets behind gene expression programming (GEP) in the mark and in expression processing.Based on analyzing and studying the elementary theory foundation of Genetic Algorithms(GA), Genetic programming(GP), we did further studies on Genetic programming(GP) and GEP's application of Data mining. This dissertation include the following several main aspects:1) Introducing the concept of symbolic regression and analyzing its characteristics and difficulties.2) Dissertating the basic principle GA and GP and their steps of solving the problem, pointing out the flaw of GA and GP.3) Presenting an method which based on the Simulated Annealing Algorithms (SAGEP) used to solve the problem of symbolic regression with arbitrary constant.4) By doing a great deal of experiment and we find out that compared with GP, gene expression programming (GEP) has much more advantages.