多智能体模型空间分析导向下的城市公园空间结构生成  

Generation of Spatial Structure of Urban Parks Based on Spatial Analysis of Agent-Based Models

作　　者：孔慧子 刘天明 廖家婕 崔柳[1] KONG Huizi;LIU Tianming;LIAO Jiajie;CUI Liu(School of Landscape Architecture,Beijing Forestry University;Bartlett Faculty of the Built Environment,University College London)

机构地区：[1]北京林业大学园林学院 [2]伦敦大学学院巴特莱特建筑学院

出　　处：《风景园林》2025年第3期73-81,共9页Landscape Architecture

摘　　要：【目的】通过分析城市兴趣点(point of interest,POI)数据,实现场地相关区域城市特征的挖掘,借助黏菌智能体模型接入区域城市特征数据,生成城市公园场地空间结构,为当下城市公园场地设计提供一种复杂系统自组织机制的空间分析方法与设计新思路。【方法】采用基于POI数据映射的黏菌智能体空间网络分析设计方法,通过多智能体模型空间信息模拟,探索城市空间功能渗透下城市公园的系统性空间功能关联,从而引导规划场地的结构生形。【结果】基于POI数据映射的黏菌智能体空间网络分析设计方法在中小场地景观设计中具备较高的可行性。多智能体模型对黏菌生长行为的模拟能够有效反映与场地结构关联的城市信息在场地空间中的渗透结果,形成带有自组织路径肌理的场地空间功能分区。【结论】多智能体模型分析借助空间算法,能有效载入场地及其关联系统空间的设计信息,通过智能体粒子模拟群体行为来形成空间关系映射,可为景观设计带来新的思考范式。[Objective]This research aims to leverage the advances of computer algorithms to provide scientific analytical methods for spatial simulation and offer a more objective approach to reflect the natural operational laws of interdependencies between elements in the real physical world for landscape design.Multi-source data characterized by high-density distribution can reflect urban features and spatial patterns at a macro scale.This research employs agent-based models(ABM)to simulate spatial information and explore the spatial function distribution of planned sites under urban spatial function penetration,thereby guiding the structural morphology of scheduled sites.[Methods]This research adopts a spatial network analysis design method using slime mold intelligence based on mapping point of interest(POI)data.Firstly,POI data and urban basic information data within the Yizhuang New City area of Beijing are obtained through the application programming interface(API)of Gaode Map to construct the spatial network foundation of the research area.The POI data are grouped into eight types:data on residential areas,hotel accommodations,road traffic,education and culture,medical care,business offices,commercial services,and green space tourism.Secondly,a slime mold agent model is constructed using the NetLogo platform,with the behavioral rules of cytoplasm agents,nuclei agents,and patch agents being defined to simulate the growth behavior of slime molds.By adjusting model parameters,the foraging paths of slime molds in the designed site are simulated,and their spatial function distribution is analyzed.The simulation environment is based on multi-source data analysis and involves setting particle growth points,food locations,and environmental obstacles.Six particle growth points represent the main entry points to the designed site,while food locations are placed at significant building entrances and preserved landscape nodes.The environmental obstacles consist of planned buildings and river channels.Finally,the model incorpora

关 键 词：风景园林 多智能体模型 兴趣点(POI)数据 空间复杂性 涌现 

分 类 号：TU986[建筑科学—城市规划与设计]