A hybrid differential evolution based on gaining‑sharing knowledge algorithm and harris hawks optimization.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Evolutionary Computation

Background:

• Differential evolution (DE) is a widely used optimization algorithm known for its simplicity and efficiency.
• However, DE often struggles to effectively balance exploration and exploitation, limiting its performance on complex problems.
• Existing DE variants require further improvements in adaptability and balance.

Purpose of the Study:

• To propose a novel hybrid algorithm, DEGH, that enhances the exploration-exploitation balance in differential evolution.
• To introduce a hybrid mutation operator and a self-adaptive crossover strategy to improve DE's performance.
• To evaluate the effectiveness of the proposed DEGH algorithm against state-of-the-art DE variants.

Main Methods:

• Developed a hybrid differential evolution algorithm (DEGH) integrating Gaining-Sharing Knowledge (GSK) and Harris Hawks Optimization (HHO).
• Constructed a hybrid mutation operator combining GSK's two-phase strategy, DE's "rand/1", and HHO's soft besiege rule.
• Implemented a novel self-adaptive crossover probability strategy where crossover probability and scaling factor jointly influence individual evolution.

Main Results:

• The proposed DEGH algorithm demonstrated a superior ability to balance exploration and exploitation.
• DEGH showed significant improvements in adapting to diverse optimization problems.
• Comparative experiments on 32 benchmark functions confirmed DEGH's superiority over eight state-of-the-art DE algorithms.

Conclusions:

• The DEGH algorithm effectively addresses the exploration-exploitation balance issue in differential evolution.
• The hybrid mutation and self-adaptive crossover strategies contribute to enhanced optimization performance.
• DEGH represents a significant advancement in differential evolution-based optimization techniques.