Efficient implicit constraint handling approaches for constrained optimization problems.

Area of Science:

• Engineering Optimization
• Computational Mathematics

Background:

• Real-world optimization problems, especially in engineering, often feature complex constraints that hinder finding feasible solutions.
• Existing methods like the boundary update (BU) technique aim to accelerate convergence by reducing the infeasible search space, but can distort the problem landscape.

Purpose of the Study:

• To extend the boundary update (BU) method for constrained optimization problems.
• To introduce novel switching mechanisms to mitigate landscape distortions caused by the BU method.
• To enhance the efficiency and effectiveness of evolutionary optimization algorithms for constrained problems.

Main Methods:

• The study implements two switching mechanisms integrated with the boundary update (BU) method.
• Mechanism 1: Optimization transitions to a BU-free state when constraint violations reach zero.
• Mechanism 2: Optimization shifts to a BU-free phase when the objective space no longer changes.

Main Results:

• The proposed method, incorporating switching mechanisms, was benchmarked against standard BU approaches on engineering and benchmark problems.
• Results demonstrate significant improvements in convergence speed.
• The enhanced method consistently finds superior solutions for constrained optimization tasks.

Conclusions:

• The developed switching mechanisms effectively address the challenges posed by the BU method's landscape distortion.
• The proposed approach offers a robust enhancement for solving constrained single- and multi-objective optimization problems.
• This research provides a valuable advancement for computational optimization in engineering and beyond.