Multi-agent Based Simulation on Residential Location Selection Based on Labor Market

LI Shao-ying,LI Xia,LIU Xiao-ping,LAO Chun-hua

DOI: https://doi.org/10.3969/j.issn.1004-6062.2013.01.006

2013-01-01

Abstract:Residential location selection is becoming a hot topic in the urban management field.There are two main research approaches including conventional regional approach and multi-agent modeling approach.The regional approaches mainly focus on statistic data involving social and economic indicators,as well as spatial data.However,the traditional "top-down" approaches are difficult in explaining the complexity of urban phenomena resulting from the decisions and behaviors of population.In contrast,the multi-agent modeling approach can simulate urban system and model the complexity of macro-processes resulting from individual behaviors at the micro-level. A variety ofmulti-agent models have been developed to simulate residential dynamics.However,these models are mainly based on the assumption of homogeneous space and do not consider the influences of spatial differentiation on economy.In fact the economic spatial configuration is the important factor that influences labor market and affects residential location selection in realistic urban environment.Therefore,the current researches have limitations on residential location selection modeling. In this paper aLabor Market-Based Model of Residential Location(LMBMRL) is proposed from the view of spatial economy,and Multi-Agent system is used to simulate the residential location selection.To model the influence of labor market,economic spatial configuration data derived from Geographical Information Systems(GIS) was incorporated into this model.As defined by spatial economic theory,the behaviors of population agents consider the influence of labor market,house cost and communication cost on individual economic utilities.Dongguan City is used for the simulation of the residential location selection by applying LMBMRL model.The simulation process reveals the effect of spatial aggregation among different population communities with certain economic spatial configuration.The simulation result demonstrates that the economic spatial configuration is likely to significantly influence the population spatial distribution which results from population residential location selection.In addition,sensitivity analysis is also carried out to assess the influence of housing cost and communication cost on population spatial distribution.The relationships among the population spatial configuration,housing cost and communication cost are detected from the simulation result.Our analysis results show that spatial structure of population is more aggregated when house cost is lower or communication cost is higher.However,as the house price rises spatial structure is less aggregated.