HMXT-GP: an information-theoretic approach to genetic programming that maintains diversity

Authors:
Henry Santosa;John Milton;Paul J. Kennedy
Affiliations:
University of Technology, Sydney, Broadway NSW, Australia;University of Technology, Sydney, Broadway NSW, Australia;University of Technology, Sydney, Broadway NSW, Australia
Venue:
Proceedings of the 2011 ACM Symposium on Applied Computing
Year:
2011

Citing 8
Cited 0

Genetic programming: on the programming of computers by means of natural selection

Genetic programming: on the programming of computers by means of natural selection
What Makes a Problem GP-Hard? Analysis of a Tunably Difficult Problem in Genetic Programming

Genetic Programming and Evolvable Machines
Limiting the Number of Fitness Cases in Genetic Programming Using Statistics

PPSN VII Proceedings of the 7th International Conference on Parallel Problem Solving from Nature
Ecological Model Selection via Evolutionary Computation and Information Theory

Genetic Programming and Evolvable Machines
Ensemble selection for evolutionary learning using information theory and price's theorem

Proceedings of the 8th annual conference on Genetic and evolutionary computation
Programming collective intelligence

Programming collective intelligence
Static and dynamic selection thresholds governing the accumulation of information in genetic algorithms using ranked populations

Evolutionary Computation
A Field Guide to Genetic Programming

A Field Guide to Genetic Programming

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper applies a recent information--theoretic approach to controlling Genetic Algorithms (GAs) called HMXT to tree--based Genetic Programming (GP). HMXT, in a GA domain, requires the setting of selection thresholds in a population and the application of high levels of crossover to thoroughly mix alleles. Applying these in a tree--based GP setting is not trivial. We present results comparing HMXT--GP to Koza--style GP for varying amounts of crossover and over three different optimisation (minimisation) problems. Results show that average fitness is better with HMXT--GP because it maintains more diversity in populations, but that the minimum fitness found was better with Koza. HMXT allows straightforward tuning of population diversity and selection pressure by altering the position of the selection thresholds.