Home > Papers

 
 
Large-scale global optimization using cooperative coevolution with self-adaptive differential grouping
FANG Wei,FANG Wei *,MIN Ruigao,ZHOU Jianhong,ZHOU Jianhong,ZHOU Jianhong
Department of Computer Science and Technology, Jiangnan University, Wuxi 214122
*Correspondence author
#Submitted by
Subject:
Funding: National Key R&D Program of China(No.2017YFC1601800, 2017YFC1601000), National Natural Science Foundation of China(No.61673194), Key Research and Development Program of Jiangsu Province, China(No.BE2017630), the Postdoctoral Science Foundation of China(No.2014M560390)
Opened online:14 December 2018
Accepted by: none
Citation: FANG Wei,FANG Wei,MIN Ruigao.Large-scale global optimization using cooperative coevolution with self-adaptive differential grouping[OL]. [14 December 2018] http://en.paper.edu.cn/en_releasepaper/content/4746608
 
 
Cooperative co-evolution (CC) is a popular evolutionary computation approach that can divide a large problem into a set of smaller sub-problems and solve them independently. CC has been an important divide-and-conquer algorithm for large-scale global optimization (LSGO) problems. Identification of variable interactions is the main challenge in CC to decompose the LSGO problems. Differential Grouping (DG) is a competitive variable grouping algorithm that can address the non-separable components of a continuous problem. As an improved version of DG, Global Differential Grouping (GDG) addresses the drawbacks of DG which are variables interactions missing during grouping and grouping performance sensitive to the threshold. In this paper, a Self-adaptive Differential Grouping (SDG) based on GDG is proposed in order to further improve the grouping accuracy on the CEC'2010 LSGO benchmark suite. 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 in the decision space. A self-adaptive pyramid allocation (SPA) strategy that can allocate different computational resource to subcomponents is also studied in this paper. The proposed algorithm, where SDG and SPA working with the optimizer $SaNSDE$ (CCSPA-SDG), is used to optimize the CEC'2010 LSGO benchmark suite. Experimental results show that SDG achieved ideal decomposition of the variables for all the CEC'2010 LSGO benchmark functions. The optimization performance of CCSPA-SDG also outperforms the state-of-the-art results.
Keywords:global optimization; large-scale global optimization; differential grouping; cooperative co-evolution; problem decomposition
 
 
 

For this paper

  • PDF (0B)
  • ● Revision 0   
  • ● Print this paper
  • ● Recommend this paper to a friend
  • ● Add to my favorite list

    Saved Papers

    Please enter a name for this paper to be shown in your personalized Saved Papers list

Tags

Add yours

Related Papers

  • Other similar papers

Statistics

PDF Downloaded 107
Bookmarked 0
Recommend 0
Comments Array
Submit your papers