随着城市化进程持续加快，城市规模不断扩大，人口快速增加，城市进入急剧扩张阶段。而城市郊区化作为快速城市化的一个重要阶段，中国的城市化发展已普遍进入这一阶段，城市发展由以往的单位主导、土地混合利用、职居接近型的规划建设模式逐渐向以市场主导、功能分区、职居分离型的规划建设模式转变。然而郊区化过程中的居住与就业不同步性、郊区新建居住区职能过于单一等原因，再加上我国城市多数以单中心布局为主，城市发展以“摊大饼”的形式扩张，致使职居分离、空间错位等现象在诸多大城市出现，造成城市居民通勤距离和时间明显增加，加重了城市交通压力，形成潮汐交通，使大城市交通拥堵无法得到治理；同时也造成城市交通能源消耗快速增加，致使交通碳排放逐渐成为城市碳排放的重要来源，加大城市碳减排难度。北京市作为中国典型的单中心布局大城市，对在后奥运时代背景下如何解决城市交通拥堵，缓解职居分离现状，缩短通勤距离和时间，减少交通碳排放成为北京市发展的重要任务。本研究针对北京职居关系改善和交通碳减排的迫切需求，按照“现状评估——实证分析——模拟优选”的思路，开展定量化分析寻找北京市职居关系失衡以及交通碳排放增长的主要原因，通过深入探讨北京市空间结构、居民通勤行为以及交通碳排放三者之间内在关系，提出一种基于职居关系调整视角的微观多智能体建模（Agent-based Modeling（ABM））耦合系统动力学（System dynamics（SD））的城市交通碳排放系统智能仿真模型，通过对北京市职居空间关系情景模拟以及多情景政策方案的综合智能仿真，给出最优选方案组合，研究结果为改善城市职居空间关系实现城市职居平衡提供思路，同时对科学制定城市职居关系改善政策措施和绿色低碳交通发展战略均具有一定的理论价值和实践意义。本研究取得以下主要成果： （1）相比2010年，2014年北京市通勤出行总量、通勤出行距离以及通勤出行时间均呈现增加趋势；而从不同尺度对北京市2010年和2014年职居分离状况分析发现，相比2010年，2014年北京市全市范围内的职居关系失衡状况有所加剧，中心城区（东城和西城）仍处于明显的职居关系失衡状态；六环内多数地区因无法提供与工作岗位数量相匹配的住房数量，从而导致职居关系失衡；同时北京市城区范围内（五环内）处于严重的因产业太过集中而住房供应不足导致的职居关系失衡，近郊区（五环和六环之间）在郊区城市化的影响下部分区域形成了以产业发展为主而住房建设不足或以大型居住区建设为主而就业岗位不足的严重的职居关系失衡状态。 （2）基于城市交通大数据（出租车GPS数据）从时空尺度上对北京市居民出行特征分析发现北京市居民出行需求和高强度出行主要集中在六环以内，工作日居民长时间（60min）和长距离（超过10km）出行主要来自六环外及周边新城（平谷、密云和怀柔），而短时间（10min）和短距离出行（小于5km）主要来自五环内区域；进一步提出一种改进的“圆圈弦图”方法对北京市居民出行时空模式进行分析，北京市区早晚高峰时间段出入量最大的地区主要是以天安门广场为中心，同时向南向北至四环的地区，其中包括北京站、北京南站和北京西站等在内的诸多交通站点；随着距离市区越来越远，其他区域的出租车出入量也是越来越小；最后利用二分K-Means聚类算法挖掘北京市职居空间模式，同时识别出典型居住地和就业地的通勤出行状况并对其进行空间可视化分析，北京市早高峰通勤起点主要分布在城市外围，即五环外的地区（包括回龙观、天通苑、通州等区域），早高峰通勤终点主要集中在城区中心的CBD、中关村以及经济开发区等产业聚集区，而晚高峰通勤状况与早高峰大致相反。不过早上从居住地到就业地的出行流和方向是更加集中和直接的，而到晚上从就业地到居住地的出行流和方向则呈现出多样化和迂回的特征。 （3）利用DEA方法建立通勤出行与职居空间关系评价模型，识别出影响两者之间协调发展的主要因素。评估结果表明职居分离是影响通勤出行的首要因素，直接影响着居民通勤时间，造成通勤时间和距离的增加，加重城市交通系统负荷；进一步分析了职居平衡对通勤交通的影响程度，研究表明相比2010年，2014年有多个区域协调度有所改善，同时也有如门头沟、大兴等区域协调度出现下降。通过对非有效DMUs进行投影分析发现北京市区域均存在不同程度输入冗余，其中在2010年DEA非有效的昌平调整量最多，平均通勤时间、平均通勤距离和过度通勤分别需减少27.2 min、10190 m和0.536；在2014年DEA非有效的大兴区调整量最多，平均通勤时间、平均通勤距离和过度通勤分别需减少20.1 min、8583.7 m和0.319；而从输出目标值看，增加区域就业密度，即通过调整各区域就业岗位数，能有效减少平均通勤时间和过度通勤，相比中心城区，近郊和远郊区的通勤交通指标能有较大幅度的改善。 （4）基于宏观视角对北京市2005-2014年交通出行碳排放量核算发现，北京市交通出行碳排放量巨大且增长迅速，年均增长6.6%，2014年碳排放量达1 065.9万吨；且通勤出行碳排放量比例呈不断上升趋势，2014年达到50.09%。在各种交通出行方式中，小汽车出行产生的碳排放量达到75%-80%，是北京交通出行碳排放产生的主要来源。进一步从微观视角利用北京市居民通勤出行调查数据对个人和家庭的通勤出行碳排放量进行了测算，结果表明北京市个人和家庭通勤出行碳排放空间特征主要以城市环路进行分隔，北京市主城区内的各环路人均和户均通勤出行碳排放量均呈现逐渐增长趋势，而在六环以外，人均和户均通勤出行碳排放量呈逐渐下降趋势；在此基础上通过构建城市个人/家庭通勤出行碳排放的影响因素解释模型发现北京市以环路分隔的家庭区位和居民职业类型是影响通勤出行碳排放量的重要因素，而那些居住在城区内的低收入且无小汽车的家庭为低碳排放者，对于那些拥有小汽车和月收入超过 40000元的家庭对通勤出行碳排放量影响较大，是造成北京市碳排放量增加的主要因素。 （5）基于职居关系调整视角利用Anylogic仿真平台构建了多智能体（ABM）耦合系统动力学（SD）的北京市交通碳排放系统智能仿真模型。该模型的ABM模块中主要包含四种Agent类型，分别为政府Agent、居民Agents、企业Agents和环境Agent，政府Agent通过制定相关政策确定政策实施后，来模拟仿真微观智能体对各政策的适应行为，进一步耦合至SD模型，构建ABM与SD耦合系统，包括人口–经济–交通子系统、经济–职居平衡–交通子系统以及交通–环境子系统；首先对居民Agents和企业Agents的空间区位进行模拟和选择，在此基础上利用构建的ABM耦合SD模型对不同情景方案交通碳减排效果进行智能仿真，从分析结果来看，通过单独实施某一政策措施虽然能削减部分交通碳排放量，但交通碳排放量仍处于增长态势，而根据北京市未来可能发展的情景，设定了三种情景方案（基准方案情景、规划方案情景和综合方案情景），其中在综合方案情景下根据政策措施减排效果进行组合，进而利用耦合系统对其进行复合动态模拟，结果表明综合方案情景能带来碳排放强度比2005年下降25%-55%的效果，同时在综合方案组合中各组合方案的环境经济效益均要高于单一政策措施，进一步以碳排放强度比2005年下降65%为目标对不同组合方案（ZHL、ZHQ和ZHB）进行模拟，并将研究目标年延长至2050年，由分析结果发现，ZHB方案的碳减排效果和环境经济效益均优于其他两种组合方案，因此，从长远目标来看综合方案情景中ZHB方案为最优发展路径，研究结果为北京市实现低碳交通提供决策参考和研究支持。

With the urbanization process continues to accelerate, the scale of the city continues to expand, the population increases rapidly, and the city enters a stage of rapid expansion. Urban suburbanization is an important stage of rapid urbanization, and urbanization development in China has gradually entered this stage. Urban development is gradually changing from construction mode of unit-led, land-mixed utilization, and jobs-housing proximity to construction mode of market-led, functionally divided, and jobs-housing separation. Due to the inconsistency of residence and employment in the process of suburbanization, and the singleness of functions of new residential areas in the suburbs, and most of the cities in China are mainly single-center layouts, urban development is expanding in the form of “spreading cakes”. The phenomenon of jobs-housing separation and spatial dislocations have occurred in large cities, resulting in a significant increase in the commuting distance and time of urban residents, increasing the pressure on urban traffic and forming tidal traffic, and making the traffic congestion in big cities impossible to be solved. At the same time, it also causes a rapid increase in urban transportation energy consumption, resulting in traffic carbon emissions becoming an important source of urban carbon emissions, increasing the difficulty of urban carbon emission reduction. As a typical single-center layout city in China, how to solve urban traffic congestion, to relieve the current situation of jobs-housing separation, shorten commuting time and distance, and to reduce traffic carbon emissions, which is an important task for Beijing's development in the background of the post-Olympic era. This study focuses on the urgent need of improving urban jobs-housing relationship and reducing traffic carbon emissions in Beijing. According to the idea of “current assessment - empirical analysis - simulation and optimization”, this study conducts quantitative analysis to find the main reasons for the imbalance of jobs-housing relationship and the growth of carbon emissions of Beijing. Further, this paper explores in depth the intrinsic relationship between urban spatial structure, residents' commuting behavior and urban traffic carbon emissions in Beijing. We propose a complex system model that couples multi-agent based models (ABM) and system dynamics (SD) models to simulate the system of urban traffic carbon emission based on the perspective of jobs-housing relationship adjustment. Through the simulation of jobs-housing relationship and the comprehensive intelligent simulation of multi-scenario policy schemes in Beijing, the most optimal combination of schemes is given. The research results provide ideas for improving the spatial relationship of jobs housing and achieving the balance of jobs housing. At the same time, they have certain theoretical and practical significance for scientifically formulating policy measures for improving jobs-housing relationship and green low-carbon transportation development strategies. The main research contents and conclusions were as follows: (1) According to the data of the fourth and fifth comprehensive survey on urban traffic in Beijing, the total amount of commuting trips, commuting distance and time showed an increasing trend compared to 2010. While the analysis of jobs-housing separation in Beijing from different scales found that, compared with 2010, imbalance of jobs-housing relationship in Beijing in 2014 has intensified. The central city (Dongcheng and Xicheng) is still in an unbalanced state of jobs-housing relationship. Most areas in the Sixth Ring Road are unable to provide the number of houses that match the number of jobs, resulting in an imbalance in jobs-housing relationships. At the same time, there is a serious imbalance in the jobs-housing relationship caused by the over-concentration of the industry and the lack of housing supply within the urban area of Beijing (in the 5th Ring Road). Under the influence of urbanization in the suburbs (between the 5th Ring and the 6th Ring), some areas have formed a serious imbalance of jobs-housing relationship with industrial development as the mainstay and inadequate housing construction or the construction of large-scale residential areas and insufficient employment. (2) The spatial and temporal characteristics of Beijing residents' travel are analyzed from time and space scales based on the urban traffic data (taxi GPS data). It found that residents' travel demand and high-intensity travel are mainly concentrated within the 6th Ring Road of Beijing. The long-time (60 minutes) and long-distance (more than 10 km) travel of Beijing residents on working days mainly comes from outside the 6th Ring Road and the surrounding new towns (Pinggu, Miyun and Huairou). Short time (10 minutes) and short distance travel (less than 5 km) mainly come from the area within the 5th Ring Road. Furthermore, a new adapted chord diagram plot is proposed to achieve the spatial-temporal scale visualization of taxi trajectory origin-destination (OD) flows. The method can characterize the volume, direction and properties of OD flows in multiple spatial-temporal scales; it is implemented using a circular visualization package in R (circlize). The areas with the largest traffic volume during the morning and evening peak hours in Beijing are mainly centered on Tiananmen Square and heading south to north to the Fourth Ring Road, including Beijing Railway Station, Beijing South Railway Station, Beijing West Railway Station and many other traffic stations. With the distance from the urban area getting farther and farther, taxi traffic in other areas is becoming smaller and smaller. Finally, the two-point K-Means clustering algorithm is used to explore the spatial pattern of jobs-housing in Beijing, and the commuting status of typical residential and employment areas is identified and spatially analyzed. The commuting starting point of morning peak is mainly distributed over the periphery of Beijing, which is the area outside the Fifth Ring Road (including Huilongguan, Tiantongyuan, Tongzhou, etc.). The morning peak commuting terminal is mainly concentrated in the CBD, Zhongguancun and economic development zones in the urban center. The situation of evening peak commute is roughly opposite to the morning peak, but there are diversified features for the commute terminal. However, the travel flow and direction from the place of residence to the place of employment in the morning is more concentrated and direct, while the travel flow and direction from the place of employment to the place of residence in the evening are diversified and circuitous. (3) The DEA method is used to establish the evaluation model of the relationship between jobs-housing spatial distribution and commuting travel, and identify the main factors that affect the coordinated development of these. The evaluation results show that jobs-housing separation is the primary factor affecting commuting travel, which directly affects residents' commuting time, causing the increase of commuting time and distance and aggravating the load of urban traffic system. Furthermore, the impact of jobs-housing balance on commuting traffic is analyzed. The study shows that compared with 2010, there are several regional coordination degrees improved in 2014, as well as regional coordination such as Mentougou District and Daxing District. At the same time, some regional coordination degrees have declined, such as Mentougou District and Daxing District. Through the projection analysis of the non-effective DMUs, it found that there are different degrees of input redundancy in Beijing. The Changping area with the DEA non-effective in 2010 has the most amount of adjustment, and the average commute time, average commute distance and excessive commuting need to be reduced by 27.2 min, 10190 m and 0.536, respectively. In 2014, the most amount of adjustment for DEA non-effective was in Daxing District. The average commuting time, average commuting distance and excessive commuting need to be reduced by 20.1 min, 8583.7 m and 0.319, respectively. From the output target value, increasing regional employment density, that is, adjusting the number of employment posts in each region can effectively reduce the average commuting time and excessive commuting. Compared with the central urban area, the commuting traffic indicators in suburbs and suburbs can be greatly improved. (4) A bottom-up approach from a macro perspective was used to analyze the changing trends of carbon emissions. The results show that traffic carbon emissions of Beijing has a huge carbon emissions and rapid growth, with an average annual growth rate of 6.6%. In 2014, carbon emissions reached 1065.9×104 t; and commuting-related CO2 emissions in Beijing in recent years have presented an increasing trend, the proportion of carbon emissions reached 50.09%. According to the proportion of carbon emissions from various modes of transportation, CO2 emissions generated by car journeys attained 75%-80%, this being the main source of transport-related CO2 emissions in Beijing. Further, the carbon emissions of individual and family commuting trips are measured and analyzed from the micro perspective, which is based on the survey data of Beijing residents' commuting trips. The results show that the carbon emission characteristics of individual and family commuting in Beijing are mainly separated by urban loops. Average individual/household commuting-related CO2 emissions inside the main urban areas showed a gradual trend of growth. Outside the 6th Ring Road, average individual/household commuting-related CO2 emissions recorded a gradual decline. A commuting-related CO2 emission model based on Tobit models from the microscopic perspective was constructed to explore the main factors affecting commuting-related CO2 emissions of individuals/households. Findings identified that household location separated by Ring Roads and the occupation type of residents were important factors affecting CO2 emissions. Commuters with access to a car, those with a higher income, and those located in the outer regions of the main urban areas produced more CO2 emissions; commuters with lower incomes, no car access, and those located in the inner regions of the main urban areas produced fewer emissions. For commuters with car availability and a monthly income of more than 40,000 CNY, these households are the major factor contributing to the increase in commuter-related CO2 emissions. (5) From the perspective of jobs-housing relationship adjustment, we propose a complex system model that couples ABM and SD models to simulate carbon emission system from urban transport of Beijing by using Anylogic simulation platform. The ABM module of the model mainly contains four types of agents, namely government Agent, resident Agent, enterprise Agent and environment Agent. The government agent simulates the adaptation behavior of micro-agents to policies by formulating relevant policies and determining the implementation of policies. The government agent determines the implementation of the policy by formulating relevant policies, and then simulates the adaptation behavior of the micro-intelligence to each policy. Further coupling to the SD model and constructing a coupled system of ABM and SD, including population - economic - transportation subsystem, economic - jobs-housing balance - transportation subsystem and transportation - environment subsystem. Firstly, the spatial location of resident Agents and enterprise Agents is simulated and selected. On this basis, we are used the ABM coupled SD model to intelligently simulate the effect of carbon emission reduction from urban transport under different scenarios. From the analysis results, although the implementation of a single policy and measure can reduce some of the traffic carbon emissions, but the traffic carbon emissions are still growing, and there is no peak in the traffic carbon emissions. According to the possible future development of Beijing, three scenarios for transport carbon emissions reduction were designed, which are Business as usual (BAU) scenario, plan scenario and integrated scenario, respectively. According to the carbon emissions reduction effects of the single policy measures form eight combinations of the integrated scenario, and use the coupled system to carry out a composite dynamic simulation. The results show that the carbon intensity can reduce by 25%-55% compared with 2005 under the integrated scenario. At the same time, the environmental economic benefits of each combination scheme in the integrated scenario are higher than the single measure, and the environmental and economic benefits of the ZHB scheme are the best. Further, this paper aims to reduce carbon intensity by 65% compared with 2005, and simulate different integrated schemes (ZHL、ZHQ and ZHB) and extend the research target year to 2050. According to the analysis results, the effect of carbon emission reduction and environmental economic benefit of the ZHB scheme are better than the other two schemes. Therefore, the ZHB scheme is the optimal development path in the integrated scenario from the long-term goal, and the research results provide decision-making reference and research support for Beijing to achieve low-carbon transportation.