A multi-objective particle swarm optimization algorithm with two-stage archive maintenance and auxiliary archive...

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Swarm Intelligence

Background:

• The external archive in Multi-objective Particle Swarm Optimization (MOPSO) is critical for balancing convergence and diversity.
• Ineffective archive maintenance can result in poor solution distribution, inaccurate convergence, and local optima trapping.

Purpose of the Study:

• To propose TAMOPSO, a novel algorithm addressing limitations in MOPSO archive maintenance.
• To enhance Pareto-front coverage, solution distribution, and convergence accuracy in MOPSO.

Main Methods:

• TAMOPSO employs a two-stage archive maintenance strategy using adaptive grids and dynamic boundary expansion.
• It integrates angle-based diversity and dual-distance convergence metrics with adaptive selection preferences.
• A bounded auxiliary archive and stagnation detection-based particle reconstruction are introduced for improved particle guidance and exploration.

Main Results:

• TAMOPSO demonstrates superior performance in balancing convergence and diversity compared to existing MOPSO methods.
• The algorithm effectively improves solution distribution and Pareto-front coverage.
• Enhanced particle guidance and global exploration capabilities were observed.

Conclusions:

• TAMOPSO offers a significant advancement in multi-objective particle swarm optimization.
• The proposed archive maintenance and particle guidance strategies effectively overcome common MOPSO limitations.
• TAMOPSO consistently produces higher-quality approximation sets across benchmark tests.