Lossless fitness inheritance in genetic algorithms for decision trees

TL;DR

This paper introduces a lossless fitness inheritance method for genetic algorithms evolving decision trees, enabling efficient fitness computation by reusing information across similar trees, supported by theoretical analysis and experiments.

Contribution

It presents a novel lossless fitness inheritance technique that leverages decision tree structure to significantly speed up genetic algorithm evaluations.

Findings

Substantial expected speed-up demonstrated on case problems

Lossless inheritance exploits divide-and-conquer nature of decision trees

Theoretical results confirmed by experimental evidence

Abstract

When genetic algorithms are used to evolve decision trees, key tree quality parameters can be recursively computed and re-used across generations of partially similar decision trees. Simply storing instance indices at leaves is enough for fitness to be piecewise computed in a lossless fashion. We show the derivation of the (substantial) expected speed-up on two bounding case problems and trace the attractive property of lossless fitness inheritance to the divide-and-conquer nature of decision trees. The theoretical results are supported by experimental evidence.