A hierarchical chain-based Archimedes optimization algorithm.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• The Archimedes optimization algorithm (AOA) is a metaheuristic known for its efficiency and minimal parameters.
• Canonical AOA suffers from slow convergence and susceptibility to local optima due to uniform agent treatment.

Purpose of the Study:

• To address the limitations of the standard Archimedes optimization algorithm (AOA).
• To propose an improved hierarchical chain-based AOA (HCAOA) for enhanced optimization performance.

Main Methods:

• Developed HCAOA, a novel algorithm that stratifies agents based on their performance levels.
• Implemented an orthogonal learning mechanism with refraction opposition for optimal individuals.
• Integrated Levy flight-based Archimedes spiral mechanism for superior individuals and applied conventional AOA for general individuals.
• Incorporated a multi-strategy boundary processing mechanism to boost population diversity.

Main Results:

• HCAOA demonstrated superior performance compared to the standard AOA and other advanced competitor algorithms on the CEC 2017 test suite.
• The proposed HCAOA achieved competitive optimization results on four complex engineering design problems.
• Experimental outcomes validate the effectiveness of HCAOA in overcoming local optima and accelerating convergence.

Conclusions:

• HCAOA effectively enhances the exploration and exploitation capabilities of the Archimedes optimization algorithm.
• The hierarchical approach and multi-strategy boundary processing significantly improve optimization speed and population diversity.
• HCAOA presents a robust and effective solution for both theoretical optimization tasks and practical engineering challenges.