A cluster-based differential evolution with self-adaptive strategy for multimodal optimization.

Area of Science:

• Computational intelligence
• Optimization algorithms
• Evolutionary computation

Background:

• Multimodal optimization is complex, requiring algorithms to find multiple global and local optima.
• Existing single-objective optimization methods are insufficient for multimodal landscapes.

Purpose of the Study:

• To develop and evaluate novel differential evolution (DE) algorithms for multimodal optimization.
• To enhance the ability of DE to locate diverse optima within complex search spaces.

Main Methods:

• A cluster-based multipopulation strategy divides the population into subpopulations for diverse search.
• Self-adaptive parameter control is integrated to improve the search capabilities of DE.
• The proposed techniques are applied to Crowding DE (CDE) and Species-based DE (SDE), creating self-CCDE and self-CSDE.

Main Results:

• The proposed self-CCDE and self-CSDE algorithms were tested on benchmark functions.
• Performance was compared against several state-of-the-art multimodal optimization algorithms.
• Experimental results validated the effectiveness and efficiency of the new multipopulation and self-adaptive strategies.

Conclusions:

• The developed multipopulation strategy and self-adaptive parameter control significantly improve DE for multimodal optimization.
• The proposed algorithms consistently achieved top rankings against competing methods.
• This research offers a robust solution for effectively addressing challenging multimodal optimization tasks.