The Effect of Multi-Generational Selection in Geometric Semantic Genetic Programming

TL;DR

This paper introduces a multi-generational selection scheme in Geometric Semantic Genetic Programming that leverages stored historical populations to improve search performance without additional costs.

Contribution

It proposes a novel multi-generational selection method in GSGP that utilizes older populations, enhancing performance efficiently.

Findings

Using older generations improves GSGP performance.

The method requires no additional computational cost.

It demonstrates the value of historical population data in evolutionary search.

Abstract

Among the evolutionary methods, one that is quite prominent is Genetic Programming, and, in recent years, a variant called Geometric Semantic Genetic Programming (GSGP) has shown to be successfully applicable to many real-world problems. Due to a peculiarity in its implementation, GSGP needs to store all the evolutionary history, i.e., all populations from the first one. We exploit this stored information to define a multi-generational selection scheme that is able to use individuals from older populations. We show that a limited ability to use "old" generations is actually useful for the search process, thus showing a zero-cost way of improving the performances of GSGP.