A Multiobjective Framework for Many-Objective Optimization.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Evolutionary Computation

Background:

• Many-objective optimization problems (MaOPs) struggle with maintaining convergence and diversity due to high-dimensional objective spaces.
• Existing algorithms often fail to balance these critical aspects effectively in complex optimization landscapes.

Purpose of the Study:

• To propose a novel multiobjective framework (Mo4Ma) for addressing challenges in many-objective optimization.
• To enhance the diversity and convergence of solutions in high-dimensional objective spaces.
• To develop a generic framework compatible with various evolutionary computation algorithms.

Main Methods:

• The Mo4Ma framework transforms many objectives into two indicative objectives: convergence and diversity.
• A clustering-based sequential selection strategy is employed in the transformed space to guide the evolutionary search.
• The framework is integrated with Differential Evolution (DE), resulting in the Mo4Ma-DE algorithm.

Main Results:

• The Mo4Ma-DE algorithm successfully obtains well-converged and widely distributed Pareto solutions for MaOPs.
• Experimental results demonstrate the algorithm's strong competitiveness against seven state-of-the-art MaOP algorithms.
• The proposed method shows generally better performance in solving many-objective optimization problems.

Conclusions:

• The Mo4Ma framework effectively transforms many-objective problems into a manageable multiobjective space.
• The Mo4Ma-DE algorithm offers a robust and competitive approach for tackling MaOPs.
• The proposed method provides a promising direction for future research in many-objective optimization.