Research On Urban Expansion Based On Ant Colony Algorithm

Urban expansion is a complex process,which involves the transformation of various types of land use into urban construction land,and results in the enlargement of city scale.The economic development has accelerated the process of urbanization.The problems related to the simulation and the prediction of urban expansion have become the hot spots of research.Cellular Automaton(CA)is a "top-down" dynamic simulation tool.CA uses simple systems to simulate complex structure,which makes it become more flexible and effective than general models.CA has very strong advantages in simulating urban expansion.The core of CA is the determination of conversion rules.Most of the current conversion rules are the implicit expressions of mathematical formulas.When the study area is more complex,it is difficult to express it.In order to describe the complex relationship in urban expansion more objectively and accurately,this paper used Ant Colony Optimization(ACO)algorithm to obtain the effective conversion rules of geographic CA instead of the expressions of mathematical formulas.This paper reviewed the basic theories related to urban expansion in detail and studied the theories of CA model and ACO algorithm,which provide theoretical support for subsequent case studies.This study selected the main city area of Changchun city as the study area,and Landsat TM images acquired in 2005,2010 and 2015 as the data sources.In addition,DEM data and GIS space vector data were selected as the secondary data source for this study.First,land use types in the study area were classified and the classification results in each period were obtained based on Remote Sensing(RS)and Geographic Information System(GIS)technology.Second,a rule mining model was established based on ACO algorithm to obtain the effective rules of affecting urban expansion.Third,these rules were applied in CA model.The CA model of GeoSOS-FLUS was used to simulate the data from the year 2005 to 2010 and from the year 2010 to 2015 respectively.Finally,the urban expansions of the study area in2020 and 2025 were predicted by CA model.The main research contents and conclusions are as follows:(1)Based on the remote sensing image data acquired in 2005,2010 and 2015,supervised classification method was used to acquire the classification data of land use for three periods.Based on ArcGIS analysis tool,the driving factor data required for the experiment were obtained.(2)The mining model of the conversion rules between different types of land use in the process of urban expansion was established based on ant colony algorithm.Somereliable urban expansion rules were excavated for the simulation and the prediction of CA.(3)GeoSOS-FLUS model was used to simulate the urban expansions from the year 2005 to 2010 and from the year 2010 to 2015 respectively.On the basis of the 2005 land use data,Simulating 2010 urban expansion and obtaining the kappa coefficient of0.93,on the basis of the 2010 land use data,Simulating 2015 urban expansion and obtaining the kappa coefficient of 0.83.Model simulation accuracy conforms to requirements.FLUS model was used to predict the urban expansion of Changchun city in 2020 and 2025.The simulation results can provide some references and data support in the macro decision of land planning for relevant departments.(4)All kinds of the statistical data collected from the year 2000 to 2015,combined with the index data which include urban expansion speed,intensity,elastic coefficient and compactness,were used to analyze the spatial expansion pattern of Changchun city from multiple perspectives.The conclusions obtained are as follows.Changchun city was at the stage of rapid urbanization and the stability of urban spatial form decreased from the year 2000 to 2015.In addition,the spatial form tended to be simple,and the irregular degree of space form increased.