Title: Large-scale global optimisation using cooperative co-evolution with self-adaptive differential grouping

Authors: Wei Fang; Ruigao Min; Quan Wang

Addresses: Jiangsu Provincial Engineering Laboratory of Pattern Recognition and Computational Intelligence, Department of Computer Science and Technology, Jiangnan University, Wuxi, China ' Department of Computer Science and Technology, Jiangnan University, Wuxi, China ' Wuxi SensingNet Industrialization Research Institute, Wuxi, China

Abstract: Cooperative co-evolution (CC) provides the divide-and-conquer framework for solving large-scale global optimisation (LSGO) problems. Identification of variable interactions is the main challenge in CC. Differential grouping (DG) is a competitive approach to find the identification and Global DG (GDG) is its improvement by introducing the global information. In this paper, a self-adaptive DG (SDG) is proposed for further improving the grouping accuracy of GDG. The threshold for grouping in SDG can adjust adaptively along with the magnitude of different functions and is determined by only two points which is a randomly sampled point and its corresponding opposite point. A self-adaptive pyramid allocation (SPA) for computational resources is also studied. The proposed algorithm, with SDG, SPA, and the optimiser SaNSDE, is used to solve the CEC'2010 LSGO benchmark suite. Experimental results show that SDG achieved ideal decomposition for all the functions and the proposed algorithm obtained competitive optimisation performance.

Keywords: large-scale global optimisation; differential grouping; cooperative co-evolution; problem decomposition.

DOI: 10.1504/IJAAC.2021.111752

International Journal of Automation and Control, 2021 Vol.15 No.1, pp.58 - 77

Received: 11 Dec 2018
Accepted: 03 Apr 2019
Published online: 14 Dec 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article