Lawrence Berkeley National Laboratory

Simulating and predicting urban patterns enables evidence-based decision-making for urban planners. Given limited resources, understanding how to improve urban residential quality and rationally plan the distribution of different quality levels warrants further study. Using cellular automata (CA) and agent-based modules, this study proposes a multi-stakeholder residential quality distribution model to analyze the future quality distribution of residential buildings under different scenarios. The proposed model comprises two modules: a CA module predicting an urban functional layout model and an agent-based module predicting residential quality distribution. The CA module develops the future functional layout of the city, upon which the multiple interests of the government, developers, and residents are taken as constraints by the agent-based module to predict residential quality distribution. The proposed model was applied as a case study to Guanxian County in Shandong Province, China. Three scenario analyses were conducted: a free development scenario, a government macro-regulation scenario, and a scenario with an adjusted preference value for residential quality. The results show that the residential quality distribution model predictions make it possible to align residential quality with residents’ needs in different scenarios; hence, unreasonable distribution characteristics can be explored to develop effective improvement measures.