Program optimization by random tree sampling

  • Authors:

  • Makoto Tanji;Hitoshi Iba

  • Affiliations:

  • University of Tokyo, Tokyo, Japan;University of Tokyo, Tokyo, Japan

  • Venue:
  • Proceedings of the 11th Annual conference on Genetic and evolutionary computation
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper describes a new program evolution method named PORTS (Program Optimization by Random Tree Sampling) which is motivated by the idea of preservation and control of tree fragments. We hypothesize that to reconstruct building blocks efficiently, tree fragments of any size should be preserved into the next generation, according to their differential fitnesses. PORTS creates a new individual by sampling from the promising trees by traversing and transition between trees instead of subtree crossover and mutation. Because the size of a fragment preserved during a generation update follows a geometric distribution, merits of the method are that it is relatively easy to predict the behavior of tree fragments over time and to control sampling size, by changing a single parameter. Our experimental results on three benchmark problems show that the performance of PORTS is competitive with SGP (Simple Genetic Programming). And we observed that there is a significant difference of fragment distribution between PORTS and simple GP.