【作者】 杨励雅；

【导师】 邵春福；

【作者基本信息】 北京交通大学 ， 交通运输规划与管理， 2007， 博士

【摘要】 城市交通与土地利用之间存在复杂的互动关系。国内外实践证明，二者的协调发展，是从根源上解决城市交通问题、实现土地利用集约化的重要前提，是决定城市能否实现可持续发展的关键。本论文立足我国城市发展实际，研究城市交通与土地利用相互关系的基础理论与方法，为城市交通规划制定、交通政策分析以及城市空间结构调整提供理论基础。首先，通过分析、总结国内外关于城市交通与土地利用相互关系的研究历程、研究成果及未来研究方向，为论文研究寻找突破口。第二，分别从宏观、微观层面，分析城市交通与土地利用的互动机理，并据此提出协调我国城市交通与土地利用关系的两大基本途径，即建立公交导向的土地开发模式和以发展轨道交通为核心的城市交通结构优化。第三，针对我国城市轨道交通建设起步晚，研究先验数据有限情况下城市轨道交通沿线土地利用形态和土地价格的变化规律，提出灰色系统云-马尔可夫链以及BP神经网络-马尔可夫链的组合预测模型，并分别以南京轨道交通1号线、北京轨道交通13号线及八通线为例，验证了模型的有效性。第四，鉴于土地利用形态与交通结构分别是表征城市土地利用系统和城市交通系统的两大关键变量，以城市总出行时间最小为目标，构建了二者的组合优化模型。针对传统算法的局限性，设计了一种新的混合算法，并以重庆市主城区为例进行了计算分析。第五，立足我国城市高密度开发特征，建立了基于TOD(Transit-Oriented Development)模式的评价指标体系，构造出评价城市交通与土地利用协调关系的DEA(Data Envelopment Analysis)模型，并对1990年-2005年北京市城八区以及2004年北京、上海等16个城市的交通与土地利用协调情况进行了评价分析。本论文的主要创新性成果如下：1、提出了刻画城市交通与土地利用互动机理的微分方程模型，通过设定不同的参数，能够模拟出不同条件下二者互动的方向、程度与演化趋势，从而为把握二者相互关系的一般发展规律，寻找协调二者关系途径，提供了理论依据。2、提出了灰色系统云-马尔可夫链的组合预测模型，该模型能够在先验数据有限的情况下，较为准确地预测出城市轨道交通沿线土地利用性质的空间分布规律。3、针对传统特征价格模型在我国应用的局限性，提出BP神经网络-马尔可夫链的组合预测模型，该模型能够在先验数据有限且存在大量不确定因素的情况下，较为准确地预测出城市轨道交通沿线房地产价格的分布规律，为定量评估我国轨道交通建设对房地产价格影响提供了新思路。4、构建了城市土地利用形态与交通结构组合优化的非线性模型，并针对传统算法的局限性，综合运用传统遗传算法、模拟退火算法和动态惩罚函数法的思想，设计一种新的混合遗传算法。该算法能够克服传统遗传算法容易过早收敛以及传统模拟退火算法全局搜索能力不足的缺陷，具有较高的运算效率和求解质量。5、在城市交通与土地利用协调关系的评价中，立足我国城市高密度开发特征，提出了基于TOD模式的评价指标体系，并采用一种新的非参数统计分析方法，构造了评价二者关系的DEA模型。DEA模型不需要预先给定评价指标权重，避免了传统评价方法中人为确定权重的主观性。另外，通过模型结果分析，能够识别出影响二者协调发展的关键因素，找出导致决策单元无效的原因，从而有目的地加以改进。更多还原

【Abstract】 There is a complicated relationship between urban transportation and land use. Practice in both China and overseas countries indicated that coordinating development of urban transportation and land use is of crucial importance for solving urban transportation problems and for making land use policies. Study on the relationship between urban transportation and land use is very helpful to realize urban sustainable development. In this dissertation, some theories and methodologies about the relationship between urban transportation and land use were proposed in accordance with the special characteristics of Chinese cities. This dissertation lays a solid theoretical foundation for urban transportation planning, traffic policy analysis, and urban spatial structure alteration.Firstly, previous studies on the relationship between urban transportation and land use in China and overseas countries were reviewed to provide helpful groundwork for this dissertation. Secondly, the interaction mechanism between urban transportation and land use was studied, from both macroscopic and microcosmic angles. According to interaction mechanism analysis, two key countermeasures that harmonize urban transportation with land use were proposed. One is to employ TOD (Transit-Oriented Development) Pattern to Chinese cities; the other is to optimize urban transportation structure by developing public transit system, especially the rail transit. Thirdly, the change pattern of land type and land price along urban rail transit was studied. An integrated model based on System Cloud Grey model and Markov chain was established to forecast the differentiation of land type along urban rail transit; meanwhile, another integrated model based on BP Neural Network and Markov chain was built to forecast real estate prices along urban rail transit. Rail transit 13 and Batong in Beijing and Rail transit 1 in Nanjing were taken as examples to illustrate the application of these two models. Fourthly, an optimization model for urban land use pattern and transportation structure was established, which can minimize the total travel time under certain restrictions of land and population. A new hybrid genetic algorithm was developed to solve the model. The calculation example based on the city zone of Chongqing showed the feasibility of the proposed model and algorithm. Lastly, a model based on Data Envelopment Analysis (DEA) was developed to give a better evaluation on the relationship between urban transportation and land use in Chinese cities. A series of evaluation indexes were proposed in accordance with the high-density development characteristics to make the model suited to Chinese cities. By applying the model to Beijing and 16 other cities, the regular pattern and key influence factors of the relationship between urban transportation and land use in Chinese cities were identified.The following are the main innovations of this dissertation:(1) A model of differential equations was built to illustrate the interaction mechanism between urban transportation and land use. By giving different value to model factors, the trend of the interaction process under different conditions was simulated. Therefore, the general law of interaction between urban transportation and land use was gained, which is very helpful to give appropriate countermeasures to coordinate them.(2) An integrated model based on System Cloud Grey model and Markov chain was established to forecast the spatial distribution of land type along urban rail transit. The model takes randomness of land type change into account; moreover, the model is available when statistics are scarce.(3) In view of inapplicability of Hedonic Price Model in forecasting real estate prices along urban rail transit in Chinese cities, an integrated forecasting model based on BP Neural Network and Markov chain was established. Using this model, the distribution pattern of real estate prices along urban rail transit can be forecasted exactly even if the statistics are scarce.(4) An optimization model for urban land use pattern and transportation structure was established, and a new hybrid genetic algorithm was developed to solve the model. The hybrid genetic algorithm overcomes shortcomings of traditional genetic algorithm and simulated annealing algorithm, and can obtain high-quality results with high efficiency.(5) A model based on DEA was proposed to evaluate the relationship between urban transportation and land use in Chinese cities. The model produces more accurate results than traditional methods do because it avoids determining the index weights. Furthermore, the key factors that influence the coordinating development of urban transportation and land use can be identified by the model.更多还原