Balanced dung beetle optimization algorithm based on parameter substitution and escape strategy.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Swarm Intelligence

Background:

• The standard Dung Beetle algorithm (DBA) exhibits strong exploitation but suffers from premature convergence and parameter randomness, leading to local optima.
• An imbalance between exploration and exploitation hinders the DBA's overall performance and robustness in complex optimization tasks.

Purpose of the Study:

• To improve the optimization performance and explore the engineering application value of the Dung Beetle algorithm.
• To address the limitations of the standard DBA, specifically premature convergence and local optima issues.

Main Methods:

• Introduction of a parabolic adaptive parameter to broaden exploration and mitigate premature convergence.
• Incorporation of a Gaussian distributed phase parameter to reduce randomness and enhance exploitation.
• Integration of a Levy flight escape strategy to balance global exploration and improve solution space coverage.

Main Results:

• The proposed Balanced Dung Beetle Optimization (BDBO) algorithm demonstrated superior convergence accuracy and generalization ability compared to the standard DBA and single-strategy variants.
• BDBO achieved a 35.29% improvement in accuracy over the original DBA.
• Statistical significance of the improvements was confirmed using the Wilcoxon rank sum test.

Conclusions:

• The enhanced BDBO algorithm effectively balances exploration and exploitation, overcoming the limitations of the standard DBA.
• BDBO shows significant potential for engineering applications, particularly in photovoltaic maximum power point tracking, outperforming existing methods.