As cities continue to expand in the 21st century, there is a growing need to develop a deeper understanding of the spatial distribution of a city’s population. Policy makers are constantly seeking for better models to predict residential patterns of different income groups to identify areas with high social risks, design future infrastructure, and plan affordable housing complex. Agent-based modeling is a powerful tool for revealing system dynamics and understanding how simple individual behaviors following prescribed rules can give rise to emergent properties. This paper presents an agent-based model that focuses on examining the residential dynamics of different income groups. Agents are endowed with heterogeneous income and social capital to choose their residences under a housing market mechanism that responds to income inequality. The model builds upon the traditional urban economic theories and simulates how individuals make residential choices given the tradeoff of between job opportunities and housing price. We found that residential patterns are qualitatively different based on income and these patterns are sensitive to the changes in the total population, housing market responsiveness, and income group ratio. In particular, our model replicates how urban processes, such as gentrification and poverty concentration in the inner-city, emerge when the housing market responds to residents’ demands. The model outcomes also reveal power dynamics between income groups such that those with higher income can have more power to impact the locations of others, and sometimes even force them to settle in suboptimal choices. We suggested this model to help urban planners and policymakers to better understand the complexity of residential dynamics, but further extensions are required to make posterior predictions that can be integrated into the planning practices.
Full capstone paper here Codes here
