Highlights of Semantics in Multi-objective Genetic Programming

TL;DR

This paper introduces a new semantic-based distance approach in multi-objective genetic programming, improving solution diversity and performance across different frameworks and datasets.

Contribution

It proposes the Semantic-based Distance (SDO) method and evaluates it alongside existing semantic approaches within EMO frameworks, demonstrating enhanced results.

Findings

SDO consistently produces more non-dominated solutions.

Semantic approaches improve diversity and hypervolume.

Enhanced performance on unbalanced binary classification datasets.

Abstract

Semantics is a growing area of research in Genetic programming (GP) and refers to the behavioural output of a Genetic Programming individual when executed. This research expands upon the current understanding of semantics by proposing a new approach: Semantic-based Distance as an additional criteriOn (SDO), in the thus far, somewhat limited researched area of semantics in Multi-objective GP (MOGP). Our work included an expansive analysis of the GP in terms of performance and diversity metrics, using two additional semantic-based approaches, namely Semantic Similarity-based Crossover (SCC) and Semantic-based Crowding Distance (SCD). Each approach is integrated into two evolutionary multi-objective (EMO) frameworks: Non-dominated Sorting Genetic Algorithm II (NSGA-II) and the Strength Pareto Evolutionary Algorithm 2 (SPEA2), and along with the three semantic approaches, the canonical form of NSGA-II and SPEA2 are rigorously compared. Using highly-unbalanced binary classification datasets, we demonstrated that the newly proposed approach of SDO consistently generated more non-dominated solutions, with better diversity and improved hypervolume results.