Mother optimization algorithm: a new human-based metaheuristic approach for solving engineering optimization.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• Optimization problems are prevalent across scientific and engineering disciplines.
• Existing metaheuristic algorithms often struggle with balancing exploration and exploitation.
• Novel approaches are needed to enhance the efficiency and effectiveness of optimization techniques.

Purpose of the Study:

• Introduce a novel metaheuristic algorithm, the Mother Optimization Algorithm (MOA).
• Simulate maternal care phases (education, advice, upbringing) for optimization.
• Evaluate MOA's performance on benchmark functions and engineering problems.

Main Methods:

• Developed a mathematical model for MOA based on human mother-child interaction.
• Tested MOA on 52 benchmark functions (unimodal, multimodal, high-dimensional, fixed-dimension) and the CEC 2017 test suite.
• Applied MOA to four real-world engineering design problems.
• Conducted statistical analysis using the Wilcoxon signed-rank test.

Main Results:

• MOA showed high ability in local search (exploitation) on unimodal functions.
• MOA demonstrated strong global search (exploration) capabilities on high-dimensional multimodal functions.
• MOA effectively balanced exploration and exploitation, outperforming 12 competing metaheuristic algorithms on most objective functions.
• MOA proved effective for real-world engineering optimization problems.

Conclusions:

• The proposed Mother Optimization Algorithm (MOA) is a novel and effective metaheuristic.
• MOA exhibits a superior ability to balance exploration and exploitation in optimization.
• MOA offers statistically significant advantages over existing algorithms for various optimization tasks.