AOBLMOA: A Hybrid Biomimetic Optimization Algorithm for Numerical Optimization and Engineering Design Problems.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• The Mayfly Optimization Algorithm (MOA) is a powerful biomimetic metaheuristic but suffers from slow convergence and local optima.
• Existing optimization techniques require enhancements to address MOA's limitations in speed and solution quality.

Purpose of the Study:

• To develop an improved metaheuristic algorithm, AOBLMOA, by combining MOA, Aquila Optimizer (AO), and opposition-based learning (OBL).
• To overcome the convergence speed and local optimization issues inherent in the original MOA.

Main Methods:

• The Aquila Optimizer's flight and hunting strategies are integrated into the male and female mayfly population movements within the MOA framework.
• Opposition-based learning (OBL) strategy replaces the gene mutation behavior in the offspring mayfly populations.
• The proposed AOBLMOA algorithm is evaluated using benchmark functions, CEC2017 numerical optimization problems, and CEC2020 real-world constrained engineering design problems.

Main Results:

• AOBLMOA demonstrated superior performance across 19 benchmark functions, confirming its effectiveness as a combined strategy.
• Comparative analysis on 30 CEC2017 problems showed AOBLMOA outperforming state-of-the-art metaheuristic algorithms.
• Validation on 10 CEC2020 engineering design problems confirmed the practical applicability and effectiveness of AOBLMOA.

Conclusions:

• The AOBLMOA algorithm effectively addresses the limitations of the original MOA, offering improved convergence and global optimization capabilities.
• The hybrid approach integrating AO and OBL provides a robust and superior metaheuristic for continuous and constrained global optimization.
• AOBLMOA shows significant promise for solving complex numerical optimization and real-world engineering design challenges.