DE/isolated/1: a new mutation operator for multimodal optimization with differential evolution

作者：Otani Takahiro^*; Suzuki Reiji; Arita Takaya

来源：International Journal of Machine Learning and Cybernetics, 2013, 4(2): 99-105.

DOI：10.1007/s13042-012-0075-y

• Differential evolution
• Multimodal optimization
• Niching

This paper proposes a new variant of differential evolution for multimodal optimization termed DE/isolated/1. It generates new individuals close to an isolated individual in a current population as a niching scheme. This mechanism will evenly allocate search resources for each optimum. The proposed method was evaluated along with the existing methods through computational experiments using eight two-dimensional multimodal functions as benchmarks. Experimental results show that the proposed method shows better performance for several functions which are not effectively solved by existing algorithms.

• 出版日期2013-4