Font Size: a A A

Analysis Of City Population Distribution Prediction Based On Multi-agent And GIS

Posted on:2008-04-29Degree:DoctorType:Dissertation
Country:ChinaCandidate:N JingFull Text:PDF
GTID:1117360215450803Subject:Human Geography
Abstract/Summary:PDF Full Text Request
Sustainability in urban development has become a critical issue due to the high levels of urbanization in almost all parts of the world. Urban growth results in environmental and social problems such as air pollution, vulnerability of watersheds and green field, increased traffic congestion, and health problems. Urban planners use a variety of tools when developing strategies and plans to mitigate these problems. Traditionally these have been prescriptive tools such as geographic information systems (GIS) or descriptive tools such as computer aided drafting (CAD) software. These tools, however, have had little predictive capability.For more than 20 years, researchers have been developing modeling approaches to describe and predict urban growth. With the driving of human beings, nature, society and economics form the city system, which is typically a complex adaptive system (CAS). Analyzing city development based on CAS is one of the main ideas these days. CAS emphasizes that the mutual effect of individuals and the environment is the driving force for system evolvement and evolution. Agents should be used to represent population in city development model, simulating the interaction between people and environment. That is to say, we should describe the real world by using the frame of the real world. However, only a few studies report on the use of agent-based models as part of spatial decision support tools for population migration.City development is spatio-temporal phenomena, whose expression needs both spatial process model and spatial data model. Since GIS data model is spatial at the core, GIS is limited when dealing with process knowledge. Process model can express time and action in detail, yet it is apparently limited in dealing with spatial knowledge. Tight integration of GIS and multi-agent system can express city development accurately. Object-oriented method has accelerated the tight integration of spatial process and data model. In this paper JAVA is chosen to integrate multi-agent platform Repast and objected-oriented GIS.Migration studies have been paid further attention in the field of human geography in recent years. In this research multi-agent, GIS and Cellular Automata (CA) are integrated together, and bottom-up modeling is used to provide decision support for city development. We use Repast and Openmap as tools in the simulation and construct independent agents to interact in the environment of air quality, green field, commercial service, house price, traffic, education and medical treatment. In this way, the real world of an evolving city is simulated. Ecological footprint method has never been used in city models. This paper uses ecological footprint index in the action rules of agents, trying to discuss some sustainable development issues. The city population model established in this paper can simulate the complex dynamic change according to different policy input, and provide valid prediction of city development. The results of the study should provide insights to policy-makers involved in urban planning.This paper attempts a novel synthesis research of population migration focusing on agent and combining 7 layers of city environment data, using Beijing, China as a case. The pace of development could be visibly seen in Beijing ever since it has undertaken the Olympics of 2008. The simulation results show that the population distribution of 2010 will become well-regulated. People will concentrate between the fourth ring and the fifth ring. In the city zone population will decrease except the Dongcheng district. Central concentration and ring distribution will be the two main characteristics of city development. Government would have controlled out-of-order expansion and central decay, reducing the population density of the center, improving the environment quality, and preventing urban decentralization. In a word, the results prove that the regulating and control plan of the government will be effective.Researchers have used CA or both multi-agent and CA to predict land use change. There are also some city models that predict the dynamics of city by simulating the interaction of different kinds of agents. The synthesis analysis that take population as agents, combine GIS and multi-layer environment to simulate population migration in city, has never been performed before. So this is the innovation point of the research. In this paper we prove that it's valid to integrate multi-agent and GIS to predict city development.
Keywords/Search Tags:City Population Distribution Prediction Model, Agent, GIS, CA, Tight Integration, Population Migration
PDF Full Text Request
Related items