Social systems are inherently complex and can be represented using agent-based modelling (ABM) methods. Based on the innovative work of Thomas Schelling, ABMs are used to represent, analyze, and forecast emergent spatial-temporal dynamics of residential segregation. Segregation is modelled by representing the complex dynamics between individual agents with various socio-demographic profiles who self-organize into spatial clusters of alike individuals. Agents are typically classified into broad panethnic categories such as “Asian” or “Hispanic”, however these categories group together individuals from a very large number of countries that are ethnically and economically distinct and thus have diverse settlement patterns. Therefore, the objective of this study is to implement an ABM that simulates the spatio-temporal dynamics of segregation that emerge from interactions between incoming immigrants who are classified at two different levels of aggregation. At the aggregate level, ethnic groups are defined based on typical broad panethnic categories. At the disaggregate level, the “Asian” category is further disaggregated. The ABM is implemented to simulate processes leading to segregation in the City of Toronto and Metro Vancouver using actual geospatial and census data. The simulated spatial patterns of segregation are compared with actual census data that records the real settlement patterns of immigrants of various ethnicities. In addition, the degree of segregation is quantified and compared with the degree of segregation measured from the actual census data. Results show that both the spatial patterns of segregation and the measure of segregation are significantly influenced by the level of aggregation of the various ethnicities. The presented research has the potential to contribute to policy-planning and decision-making by assisting city planners and policymakers in mitigating persistent residential segregation.

社会系统本质上是复杂的，可以使用基于代理的建模 (ABM) 方法来表示。基于 Thomas Schelling 的创新工作，ABM 用于表示、分析和预测住宅隔离的紧急时空动态。隔离是通过表示具有各种社会人口特征的个体代理之间的复杂动态来建模的，这些个体自组织成相似个体的空间集群。代理人通常被分为广泛的泛民族类别，例如“亚洲人”或“西班牙裔”，但是这些类别将来自大量国家的个人聚集在一起，这些国家在种族和经济上各不相同，因此具有不同的定居模式。所以，本研究的目的是实施一个 ABM，模拟隔离的时空动态，这些动态来自被归类为两个不同聚集级别的入境移民之间的相互作用。在总体层面上，族群是根据典型的广泛泛族类别来定义的。在分解层面，“亚洲”类别进一步分解。实施 ABM 以使用实际地理空间和人口普查数据模拟导致多伦多市和大温哥华地区隔离的过程。将模拟的种族隔离空间模式与记录各族移民真实定居模式的实际人口普查数据进行比较。此外，隔离程度被量化并与根据实际人口普查数据测量的隔离程度进行比较。结果表明，种族隔离的空间格局和隔离措施都受到各民族聚集程度的显着影响。所呈现的研究有可能通过协助城市规划者和政策制定者减轻持续的住宅隔离来促进政策规划和决策。