Multi-objective spatially constrained clustering for regionalization with particle swarm optimization. Issue 4 (3rd April 2018)

• Record Type: Journal Article
• Authors: He, Weixiong; Ling, Haifeng; Zhang, Zhanliang; Gong, Congcong
• Is Part Of: International journal of geographical information science. Volume 32:Issue 4(2018)
• Page Start: 827
• Page End: 846
• Publication Date: 2018-04-03
• Subjects: Spatial analysis -- regionalization -- constrained clustering -- multi-objective optimization -- particle swarm optimization; Geography -- Data processing -- Periodicals; Information storage and retrieval systems -- Periodicals; Géomatique -- Périodiques; Systèmes d'information -- Périodiques; 910.285
• DOI: 10.1080/13658816.2017.1418363
• Languages: English
• ISSNs: 1365-8816

Abstract

ABSTRACT: Regionalization is an important part of the spatial analysis process, and the solution should be contiguity-constrained in each region. In general, several objectives need to be optimized in practical regionalization, such as the homogeneity of regions and the heterogeneity among regions. Therefore, multi-objective techniques are more suitable for solving regionalization problems. In this paper, we design a multi-objective particle swarm optimization algorithm for solving regionalization problems. Towards this goal, a novel particle representation for regionalization is proposed, which can be expressed in continuous space and has flexible constraints on the number of regions. In the process of optimization, a contiguous-region method is designed that satisfies the constraints and improves the efficiency. The decision solution is selected in the Pareto set based on a trade-off between the objective functions, and the number of regions can be automatically determined. The proposed method outperforms six regionalization algorithms in terms of both the number and the quality of the solutions.