Adaptive ranking mutation operator based differential evolution for constrained optimization.

Area of Science:

• Numerical Optimization
• Evolutionary Computation
• Constraint Handling

Background:

• Differential evolution (DE) is a robust evolutionary algorithm (EA) for numerical optimization.
• DE, enhanced with constraint-handling techniques, is increasingly applied to constrained optimization problems (COPs).

Purpose of the Study:

• To introduce the adaptive ranking mutation operator (ARMOR) for DE to accelerate convergence and improve feasibility in COPs.
• To enhance the performance of constrained DE (CDE) variants through adaptive solution ranking.

Main Methods:

• The proposed ARMOR operator adaptively ranks solutions based on population status: infeasible, semi-feasible, or feasible.
• Solution ranking in ARMOR considers constraint violations, transformed fitness, and objective function values.
• Selection probabilities are dynamically adjusted based on the population's situation.

Main Results:

• ARMOR integration accelerated original CDE variants across most tested benchmark functions (CEC 2006 and CEC 2010).
• ARMOR-based CDE demonstrated highly competitive performance compared to state-of-the-art evolutionary algorithms.
• The operator proved effective in speeding up the achievement of feasible solutions.

Conclusions:

• The adaptive ranking mutation operator (ARMOR) is a simple yet effective enhancement for DE in solving COPs.
• ARMOR significantly improves the convergence speed and feasibility-finding capability of CDE algorithms.
• This approach offers a promising direction for advancing evolutionary computation in complex optimization tasks.