Mengjie Zhang, Reuse of Subprograms in Genetic Programming

Reuse of Subprograms in Genetic Programming for Classification and Regression Tasks

• Overview:

  In Genetic Programming (GP), programs are usually expressed as trees. This form of GP has been successfully applied to a number of problems such as image enhancement, image recognition, and regression problems. However, the genetic programs generated in the evolutionary process usually contain a large amount of redundancy, which often leads to slow training and large space occupation. In addition, the programs in the current generation has a huge amount of redundancy with those in the previous generation, since the recombination (crossover) just swaps genetic materials between programs rather than producing new ones. From programming design point of view, this is a big waste of space.

  One the other hand, reuse of code is a very important characteristic in human programming, such as using Java and C++. So, attempts introducing reuse of code into genetic programming in GP would be very interesting.

• Tasks and Goals of this Project

  The goal of this project is to investigate ways of reusing subprograms in genetic programming. Specifically, this project will investigate:

  • Whether the reuse of code can be implemented in GP;
  • Whether reuse of code can be introduced in single genetic programs;
  • Whether reuse of code can be introduced in different genetic programs in a particular generation; and
  • Whether reuse of code can be introduced in different genetic programs across different generations;

  The methods developed will be examined on some regression, classification or similar problems.

• More information can be seen from Meng's research projects . Related Publications can be viewed from here: Meng's recent publications. Contact me if you want to get a copy of those papers, or if you want to know more detail about this project.