A guided epsilon-dominance arithmetic optimization algorithm for effective multi-objective optimization in engineering design problems

doi:10.1007/s11042-023-16633-x

2023-09-18

PRE

Djaafar Zouache *

Laith Abualigah

Farid Boumaza

DOI：10.1007/s11042-023-16633-x

原文链接

原文求助

摘要

En 中文

The Arithmetic Optimization Algorithm (AOA) was recently proposed as a solution for single-objective real continuous problems and has demonstrated superior performance in various contexts. This paper presents a multi-objective version of the algorithm to solve multi-objective problems. The MOAOA employs an archive repository to keep and retrieve the non-dominated solutions produced during optimization. The leaders are then selected from the population archive to lead the solutions of the main population toward the potential search locations. The epsilon-dominance and crowding distance strategies balance the convergence and diversity of the obtained Pareto set. To assess the effectiveness and efficiency of the proposed algorithm, it was tested on various dimensions of multi-objective benchmarks, among them: five unconstrained test functions taken from ZDT-series and four multi-objective constrained tests functions (BNH, SRN, TNK, OSY). Also, it is evaluated on four multi-objective structural design problems, such as welded beam design, speed-reduced design, disk brake design, and four-bar truss design. The algorithm is compared with three algorithms widely used in multi-objectives optimization, such as MSSA, MOEA-D, and MOGWO. The comparison results demonstrate that MOAOA outperforms all other algorithms in terms of both convergence and diversity of solutions, achieving a score of (100%,100%) in the ZDT tests, (75%,50%) in the constrained test functions, and (75%,75%) for the structural design problems.

Keyword:

Multi-objective optimization

Crowding distance

Epsilon-dominance

Arithmetic optimization algorithm

Engineering design application