资源科学 ›› 2019, Vol. 41 ›› Issue (4): 729-739.doi: 10.18402/resci.2019.04.10

• 土地资源 • 上一篇    下一篇

基于多智能体粒子群算法的松嫩平原土地利用格局优化

王越1, 宋戈2, 吕冰1   

  1. 1. 沈阳师范大学管理学院,沈阳 110034;
    2. 东北大学土地管理研究所,沈阳 110169
  • 收稿日期:2018-05-27 修回日期:2018-09-13 发布日期:2019-04-25
  • 通讯作者: 国家自然科学基金项目(41571165); 辽宁省社科规划基金项目(L17CGL007); 辽宁省教育厅科学研究项目(WQN201709)
  • 作者简介:王越,女,黑龙江哈尔滨人,讲师,硕士生导师,主要研究方向为土地利用与管理。E-mail: ywmath@163.com

Optimization of land-use pattern based on multi-agent particle swarm optimization in the Song-Nen Plain region

WANG Yue1, SONG Ge2, LV Bing1   

  1. 1.School of Management, Shenyang Normal University, Shenyang 110034, China;
    2. Institute of Land Management, Northeast University, Shenyang 110169, China
  • Received:2018-05-27 Revised:2018-09-13 Published:2019-04-25

摘要: 土地利用格局优化是实现土地资源合理配置的重要方式。本文以松嫩平原典型区域巴彦县为研究区,运用GIS和RS技术,采用Matlab编程,结合多智能体(MA)和粒子群算法(PSO),建立土地利用格局优化模型,以粮食生产、生态安全和社会经济发展为优化目标,设计政府、职能部门和个体三类智能体(Agent),并结合研究区土地利用格局优化目标的决策偏好确定其优化方案。结果表明:①基于多智能体改进粒子群优化算法建立土地利用格局优化模型,实现了研究区土地利用类型数量结构在时空上的合理匹配及其空间构型和空间组合方式合理配置,建立的土地利用格局优化模型可行;②3个优化方案中,方案1更加偏向于生态安全优化目标的实现,方案2更加偏向于社会经济发展目标的实现,方案3更加偏向于粮食生产目标的实现;在研究区土地利用格局子目标决策偏好的实现上,3组土地利用格局优化方案对实现粮食生产、生态安全、社会经济发展等优化目标的决策偏好具有显著差异;③研究区土地利用格局不同优化方案呈现不同土地利用结构,在空间布局上有显著的分异特征。本文完善和丰富了土地利用格局优化的理论基础和研究方法,可为土地利用规划提供有力的技术支撑。

关键词: 土地利用, 格局优化, 多智能体粒子群算法, 松嫩平原

Abstract: Land-use pattern is an important indicator for the evaluation of land use and useful for analyzing and explaining the spatial phenomena, processes, and mechanisms of regional land use. The main aims of this study were to establish the optimization model of land-use pattern in a typical area of the Song-Nen Plain regionBayan Countybased on GIS and remote sensing technologies, Matlab programming, and multi-agent particle swarm optimization (MA-PSO); to optimize grain production, ecological security, and socioeconomic development with three agents of government, authorities, and individuals; and to assess the optimization schemes of land-use pattern based on the decision of land-use pattern optimization subgoals. The results indicated that the optimization model of land-use pattern based on GIS and remote sensing technologies, Matlab programming, and MA-PSO model could achieve a reasonable matching of the quantitative structure of land-use types in time and space and the rational allocation of spatial configuration and combination. The optimization objective functions of scheme I, II, and III were ecological security, socioeconomic development, and improved grain production, respectively, and the three optimization schemes had significant differences in the realization of the optimization subgoals. The different optimization schemes of land-use pattern had different land-use structures and spatial distribution characteristics. The study has enriched the theoretical basis and research methods of land-use pattern optimization, which can provide powerful technical support for land-use planning.

Key words: land use, pattern optimization, multi-agent particle swarm optimization, Song-Nen Plain