Study Of Urban Environment-Oriented Traffic Zoning Based On Clustering Algorithm

With the acceleration of urbanization and the rapid increasing number of motor vehicles, traffic congestion has become one of the most serious problems that hinder the development of numerous big cities in the world. Additionally, traffic voice and vehicle emission are regarded as the main source of city environmental pollutants, and have great adverse impacts on the life standards of modern citizens. Ecological urban transport system is the evitable direction of future transportation system.Urban transport environment system is a complex unit, performing integrated functions of transportation, social, economical, environmental services. Coupled with its complexity and openness, such elements as land use, habitant, working place, and socio-economical conditions are highly related to transport system, and showing high degree of spatial imbalance and spatial difference. This situation determines that, the traffic conditions and key problems at one place may be quite different from other place. In this mean, the improvement of urban transport environment should be region-oriented, highly depended on the specific and regional conditions. Effectively dividing whole city into different parts, should be the first step of figure out region specific measurement.This study focuses on the zoning method of urban environment-oriented transport system, the main contents are as follows:At first, the necessity of urban transport environment zoning is discussed based on the characteristics of urban transport system. Zone management applications in different fields are studied and compared. It is concluded that grid management is a good way of complex system management. Based on the concept of grid management, a new method for urban transport environment zoning is brought forward in the use of grid clustering.Secondly, an environment-oriented urban traffic zoning index system is established. The impact of different elements on the whole urban traffic system, traditional traffic system evaluation models, and the ecological transport system study framework of Driving-Pressure-State-Impact-Response (DPSIR) are the main considered perspectives. The final zoning index system contains four types of indices: geographical indices, transportation facility indices, dynamic traffic parameters, and environment indices.Thirdly, a zoning-fitted clustering algorithm is designed. Since a large number of indices are involved, and various indices follow no fixed distribution characteristic, many traditional clustering methods are not good choices. However, clustering algorithm enlightened from biological immune system has great advantages at data processing, pattern recognition, fast memory and learning. In this study, a clustering method is designed to realize the transport zoning.Fourthly, a software system of urban transport environment zoning is designed. The component library of ArcGIS Engine and designed immune clustering method are integrated into the same system to realize the automatical zoning of urban transport environment system.At last, the zoning software system is applied in the case study of Beijing. The zoning results are analyzed. On the other hand, the congestion pricing policy in the central zone is tested. And the further measurements to improve the condition of Beijing transport environment are discussed.