Adaptive niching selection-based differential evolution for global optimization

doi:10.1007/s00500-022-07510-0

Adaptive niching selection-based differential evolution for global optimization

2022-10-06

1

PRE

AI

DOI：10.1007/s00500-022-07510-0

Original

Original request for help

Save

Abstract

En 中文

Niching techniques have been widely incorporated into differential evolution (DE) to improve the diversity of the search. Existing niching-based DE schemes, however, typically employ a certain niching technique during the entire evolution. Since different niching techniques possess different diversity preserving properties, employing a fixed niching technique during DE evolution may have limited performance. In this paper, we propose an adaptive niching selection-based DE for global optimization problems. In the proposed method, instead of employing a certain fixed niching technique, an adaptive niching selection scheme has been devised. In this scheme, multiple niching techniques are employed and adaptively used during the DE evolution, thus properly preserving the population diversity. Further, to appropriately facilitate the adaptive niching selection, both the fitness improvement and entropy of population resulting from niching techniques have been considered to measure their effectiveness. The performance of the proposed method has been evaluated on multi-modal and hybrid composition test functions and compared with related methods. The results show that our proposed method can deliver a satisfying performance and is competitive with state-of-the-art algorithms.

Keywords:

Differential evolution

Niching technique

Adaptive selection

Global optimization

Adaptive niching selection-based differential evolution for global optimization

Abstract

Journal

Organization