A novel differential evolution algorithm with multi-population and elites regeneration.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Evolutionary Computation

Background:

• Differential Evolution (DE) is a powerful global optimization algorithm with advantages like ease of implementation and speed.
• However, DE faces limitations in suboptimal solution exploitation and parameter tuning.
• Addressing these challenges is crucial for advancing optimization techniques.

Purpose of the Study:

• To introduce Enhanced Binary JADE (EBJADE), a novel algorithm enhancing Differential Evolution (DE).
• To improve exploitation capabilities and address parameter tuning challenges in DE.
• To present a robust optimization algorithm for complex problems.

Main Methods:

• EBJADE combines Differential Evolution (DE) with multi-population and elite regeneration strategies.
• A novel strategy perturbs the target vector using sorted vectors for enhanced exploitation.
• Multi-population approach with a rewarding subpopulation dynamically allocates mutation strategies.
• Incorporates elite sampling from Estimation of Distribution Algorithm (EDA) for solution regeneration.

Main Results:

• Experimental results on CEC2014 benchmark tests validate EBJADE's performance.
• EBJADE demonstrates strong competitiveness against existing optimization algorithms.
• The proposed algorithm shows superior performance in global optimization tasks.

Conclusions:

• EBJADE effectively overcomes limitations of standard Differential Evolution (DE).
• The novel strategies enhance solution exploitation and parameter adaptability.
• EBJADE represents a significant advancement in evolutionary optimization algorithms.