Multi-Strategy Improved Dung Beetle Optimization Algorithm and Its Applications.

Area of Science:

• Computational Intelligence
• Swarm Intelligence
• Optimization Algorithms

Background:

• The standard dung beetle optimization (DBO) algorithm, while effective, struggles with low population diversity and local optima in complex scenarios.
• Existing metaheuristic algorithms often face challenges in balancing exploration and exploitation for robust optimization.
• Addressing these limitations is crucial for advancing swarm intelligence applications.

Purpose of the Study:

• To propose and evaluate a novel multi-strategy improved dung beetle optimization algorithm (MDBO).
• To enhance the DBO algorithm's population diversity, local optima avoidance, and convergence speed.
• To validate the MDBO algorithm's performance on benchmark functions and real-world engineering problems.

Main Methods:

• Implemented Latin hypercube sampling for improved initial population distribution.
• Introduced a "Mean Differential Variation" strategy to enhance local optima evasion.
• Integrated lens imaging reverse learning with dimension-by-dimension optimization for the best solution.
• Tested MDBO against benchmark functions from CEC2017 and CEC2020.
• Applied MDBO to engineering design problems: spring, reducer, and welded beam.

Main Results:

• MDBO demonstrated significantly superior optimization accuracy, stability, and convergence speed compared to classical metaheuristics.
• Performance improvements were validated across standard benchmark test suites (CEC2017, CEC2020).
• MDBO effectively solved complex engineering design problems, showcasing its practical applicability.

Conclusions:

• The proposed MDBO algorithm offers a robust enhancement over the standard DBO.
• MDBO effectively addresses the limitations of low population diversity and local optima entrapment.
• The algorithm shows strong potential for solving complex optimization tasks in both theoretical and practical engineering domains.