An Enhanced Randomized Dung Beetle Optimizer for Global Optimization Problems

TL;DR

This paper introduces an improved version of the Dung Beetle Optimizer algorithm to better solve complex optimization problems in engineering.

Contribution

The ERDBO algorithm introduces a three-stage enhancement mechanism to improve exploration, exploitation, and convergence.

Findings

ERDBO outperforms other metaheuristic algorithms on CEC2017 benchmark functions.

The algorithm shows improved convergence rate, stability, and solution precision.

ERDBO successfully solves three engineering design problems effectively.

Abstract

The Dung Beetle Optimizer (DBO) has shown promise in solving complex optimization problems, yet it often suffers from premature convergence and limited accuracy. To overcome these limitations, this paper proposes the Enhanced Reproductive Dung Beetle Optimizer (ERDBO). The ERDBO introduces a three-stage mechanism: (1) a larval growth phase using experiential learning to enrich population diversity and improve global exploration; (2) a reproduction and nurturing phase that employs parent–offspring verification and a teaching strategy to strengthen local exploitation; and (3) a predator avoidance phase integrating Lévy flight and sinusoidal perturbations to enhance adaptability and accelerate convergence. The effectiveness of the proposed algorithm is assessed using the CEC2017 benchmark functions, where it is contrasted with several advanced metaheuristic approaches. The experimental findings highlight its advantages in terms of convergence rate, stability, and solution precision. Furthermore, the ERDBO is applied to three well-known engineering design tasks—namely the tension/compression spring, the three-bar truss, and the pressure vessel problem. The outcomes verify both its efficiency and applicability, indicating that the ERDBO provides a robust and competitive optimization framework for tackling challenging real-world engineering scenarios.

Figures