Cooperative Differential Evolution Framework for Constrained Multiobjective Optimization.

Area of Science:

• Computational Intelligence
• Evolutionary Computation
• Optimization

Background:

• Constrained multiobjective optimization problems (CMOPs) are challenging due to simultaneous optimization of multiple objectives under constraints.
• Existing algorithms often struggle to effectively handle the complexities of CMOPs.

Purpose of the Study:

• To introduce a novel cooperative differential evolution framework (CCMODE) for CMOPs.
• To extend existing single-objective optimization techniques to address multiobjective problems with constraints.

Main Methods:

• The CCMODE framework utilizes (M+1) populations: M subpopulations for single-objective optimization and an archive population for multiobjective optimization.
• Constraint handling techniques (CHTs) are adapted for the archive population to manage constraints in multiobjective settings.
• Two instantiations of CCMODE are implemented, each employing different CHTs.

Main Results:

• Experimental results on benchmark problems with 2, 3, and many objectives demonstrate CCMODE's superior performance.
• The proposed algorithm significantly outperforms state-of-the-art constrained multiobjective evolutionary algorithms.
• The effectiveness of the subpopulations within the CCMODE framework is validated.

Conclusions:

• The CCMODE framework offers an effective approach to solving constrained multiobjective optimization problems.
• CCMODE successfully leverages and extends existing single-objective optimization strategies for multiobjective applications.
• The proposed method shows significant promise for advancing research in constrained multiobjective evolutionary algorithms.