APEX-DE: Adaptive Parameter Control and Selection Strategy for Differential Evolution With Exponential Crossover.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Evolutionary Computation

Background:

• Differential evolution (DE) algorithms commonly utilize binomial crossover.
• Research has shown exponential crossover can yield better results with proper parameter control.

Purpose of the Study:

• Introduce an adaptive parameter control and selection strategy for DE with exponential crossover (APEX-DE).
• Develop a high-performance DE variant using exponential crossover.

Main Methods:

• Proposed a novel adaptive parameter control (APC) technique with automatic crossover rate generation.
• Implemented a dual-stage scale factor generation mechanism and adaptive strategy for scale parameter.
• Introduced a new selection mechanism to improve local optima escape and a redirection strategy for enhanced evolutionary potential.

Main Results:

• APEX-DE demonstrated superior performance on 88 benchmark functions.
• The algorithm achieved better results on a challenging uncrewed aerial vehicle (UAV) path-planning task.
• Experimental evaluations confirmed APEX-DE outperforms state-of-the-art algorithms.

Conclusions:

• APEX-DE provides a high-performance DE variant leveraging exponential crossover.
• The proposed adaptive strategies significantly enhance optimization capabilities.
• APEX-DE shows promise for complex real-world problems like UAV path planning.