Multi-objective artificial ecosystem optimization algorithm : a novel optimization approach based on shift-based density estimation and external archive population

Farid Boumaza

doi:10.1007/s10586-026-06568-x

Return

Multi-objective artificial ecosystem optimization algorithm : a novel optimization approach based on shift-based density estimation and external archive population

2026-09-20

PRE

Farid Boumaza *

Djaafar Zouache

DOI：10.1007/s10586-026-06568-x

Original

Original request for help

Save

Abstract

En 中文

Artificial Ecosystem-based Optimization (AEO) is a nature-inspired metaheuristic originally designed for single-objective optimization. However, its potential for multi-objective scenarios remains largely unexplored, limiting its use in complex multi-objective optimization problems (MOPs). To bridge this gap, we propose a Multi-Objective Artificial Ecosystem-based Optimization algorithm (MOAEO) that extends AEO to effectively solve multi-objective problems, including multi-objective engineering design problems (MOEDPs). The proposed MOAEO framework integrates an external archive to preserve Pareto-optimal solutions, applies shift-based density estimation (SDE) to promote diversity, and utilizes non-dominated sorting to guide elite selection and enhance convergence. We conduct extensive experiments on a diverse set of benchmark problems, including bi-objective constrained and unconstrained cases, three-objective and many-objective test functions, and real-world engineering design problems. Experimental results show that MOAEO achieves competitive convergence and diversity across multiple benchmark suites. In many instances, it delivers statistically significant improvements over existing methods, while maintaining comparable performance in more challenging scenarios. These findings highlight MOAEO’s potential as an effective approach for solving complex multi-objective optimization problems.

Keywords:

Artificial ecosystem-based optimization

Multi-objective optimization

Non-dominated sorting

Shift density estimation

Pareto front