王忠宇
大型活动是指具有一定计划性、在特定的时间和地点发生、参加人员众多并且能引起交通需求不正常增长以至对城市正常交通运行造成明显影响、需要交通管理与控制手段干预的特殊事件。随着经济社会的发展，各类大型活动举办日渐频繁，并有渐成常态化的趋势。各类大型活动所引发的异常交通需求，对于以常态交通需求为目标的城市交通系统是一个严峻的挑战。按照是否为举办大型活动而新建或改建交通基础设施来区分，可将大型活动分为两类：一类是需要新建或改建交通设施以满足大型活动期交通需求的，此类大型活动通常规模大、计划时期长，活动持续时间也较长；另一类是利用现有交通设施，主要通过各类交通管控措施确保大型活动期交通正常运行的，此类大型活动通常规模较小、计划时期较短、活动持续时间也较短。本文以减少拥挤对路网造成的影响进而确保大型活动参与者顺利出行、居民日常出行不受严重影响为目标，在现有交通设施条件下，针对日趋常态化的规模一般、近期实施的大型活动可能诱发的机动车交通拥挤，构建系统完善的拥挤管理理论与方法，对适应大型活动期交通拥挤事前预防和事中疏解的管理与控制策略进行研究。
在从交通角度定义大型活动并细化其分类的基础上，本文通过实际数据的分析，指出大型活动诱增出行需求在数量、集中性、层次性、可靠性要求等几个方面表现出与常规出行需求不同的特征，进而分析并凝练出大型活动期道路交通流具有不确定性、时空分布不均、时空波动显著、源汇区别特征明显、交通方向不均衡系数大等5个特征和运行规律。通过对决定于交通流特征的大型活动期道路交通拥挤演化的分析，得到了拥挤自核心向周边发散、拥挤沿一定方向渐进式传播、拥挤方向性明显等3个关键特征，并比较分析了其他类型大型活动期交通运行和拥挤演化的相似或相异特征。
基于系统科学的理念，本文构建了完整的大型活动期道路交通拥挤管理理论体系，明确将研究聚焦于现有道路网条件下大型活动期道路交通拥挤的事前预防和事中疏解管控策略。借鉴拥挤管理系统（CMS）思想，提出了包含拥挤识别与评估、拥挤管控策略以及方案效果评估和调整三大模块的拥挤管理理论体系及其逻辑框架，并对支撑体系的交通流理论、系统科学理论、最优化理论、控制论等几大基础理论及其指导作用进行了分析阐述，对较为关键的交通检测技术、计算机和通信技术、人工智能和智能计算技术等几方面相关技术及其应用进行了分析归纳。
大型活动期的拥挤识别与评估，评价指标十分关键，而且大型活动期各类拥挤管控策略的调用判断和效果评估很大程度上也依赖于这些评价指标。本文对其中两个关键指标——交叉口的排队长度、延误的新的获取方法进行了研究。基于多源数据获取的便利，提出了估计信号交叉口排队长度的破解初始排队长度获取难题的HCM 2000模型改进方法和不依赖于高分辨率线圈数据或多条浮动车数据的基于波动论的方法，并分别进行了实证研究。利用信号交叉口的实际数据对比分析了4种延误估计模型，研究了不同饱和度条件下各模型精确度的变化趋势及其影响因素，分析了各模型不同情形下的适用性。
为减少近期实施的大型活动引发道路交通拥挤的可能性和风险，针对拥挤沿一定方向渐进式传播、方向性明显等演化特征，本文在对网络交通变结构控制优化概念清晰界定的基础上，分别提出以复合饱和度CL'C最小为优化目标的适应大型活动期拥挤预防的路段变结构控制优化方法和与之相适应的交叉口变结构控制优化方法，针对大型活动影响范围区域内部交通、跨区交通、过区交通提出了协同辅助网络交通变结构优化的缓解交通压力的群体诱导策略，并进行了案例分析。
针对规划手段和活动举办前短时间的预防策略所不能完全顾及的大型活动期偶发性交通拥挤，本文构建了分别针对面拥挤、线拥挤、点拥挤的由局域网络管控、路径协调控制、单点控制所构成的拥挤疏解策略体系。提出了针对拥挤自核心向周边发散的基于模糊控制的小区边界交叉口截流—卸载控制方法，并设计了模糊控制器；提出了针对拥挤沿一定方向渐进式传播的基于绿波和红波理念的路径协调控制方法，重点研究了运用图解法确定相位差的方法；并利用仿真验证了各策略的实施效果。
本文通过对大型活动期道路交通运行和拥挤演化特征的认知和把握，基于系统科学理念，构筑包括基于评价指标的拥挤识别与评估、基于时空优化的拥挤预防与疏解的以精细化管理为导向的拥挤管理理论体系，丰富和完善了特殊事件条件下交通管理与控制的理论与方法。限于数据、现场实验等条件和管理理论体系本身及其实施的复杂性，研究尚有诸多不足和缺陷，特别是未能对策略效果进行现场实验检验和分析，这是后续研究的重要内容。同时，大型活动期道路交通拥挤管理平台化开发也是未来的重要发展方向。
关键词：大型活动，拥挤管理，网络交通变结构，协调控制，最优化方法
Special events are the planned events which occur at known locations and on scheduled times. As the participants are multitudinous, a special event might create an abnormal increase in travel demand, thus impacting the operation of urban transportation system, and requires traffic management and control measures. With the economic and social development, a variety of special events are organized frequently. The abnormal travel demands created by special events bring about numerous challenges to the urban transportation system, which has been built to serve normal travel demands. According to whether transport facilities need to be newly-built or rebuilt to satisfy travel demands during events, special events can be divided into two categories: 1) transport facilities need to be newly-built or rebuilt in order to meet the travel demands during special events, which are usually large-scale, and have longer planning periods and durations; 2) existing transport facilities are able to satisfactorily meet the travel demands during special events mainly through various traffic management and control measures, which are generally small-scale, and have shorter planning periods and durations. This study targets at reducing the impact of congestion on road networks, which are caused by special events, and thus ensuring a smooth traffic of special events participants and minor effects on daily trips of residents. In order to achieve the target, a complete system of congestion management theories and methods is built. The system makes use of existing transport facilities, and implements traffic management and control strategies to prevent and mitigate traffic congestion before and during special events, which are more and more normalized, general-scale and will be implemented recently.
This study first defines special events and their classification from the perspective of transportation. Then, according to the definition and actual data analyses, this study finds out the characteristics of travel demands created by special events are different from that of normal daily travel demands in the following four aspects: quantity, centrality, hierarchy, and requirement of reliability. Additionally, it summarizes five characteristics and operating rules of the road traffic flow during special events, including uncertainty, uneven distribution, temporal and spatial fluctuation, distinct difference of generation and attraction, and large uneven coefficient of traffic direction. Through analyzing the characteristics of traffic congestion evolution during special events, this study discovers three key features of congestion, including diffusion from the core to the periphery, dissemination in a certain direction gradually, and significant directionality. Also, a comparative analysis is conducted on the similar or dissimilar traffic operation and congestion evolution characteristics of other types of special events.
This study proposes a complete road networks congestion management theory system during special events based on the concept of system science. The study clearly aims at the management and control strategies on traffic congestion preventing and mitigating before and during special events under the existing road networks. Being inspired by the Congestion Management System (CMS), this study introduces the congestion management theory system and describes its logical framework during special events. The system contains congestion identification and assessment module, congestion management and control strategies module, and program effectiveness evaluation and adjustment module. Meanwhile, basic theories which support the system are analyzed and discussed. Furthermore, major critical technologies adopted in this study are analyzed and summarized.
For congestion identification and assessment during special events, evaluation measures are critical, because the calling judgment and effectiveness evaluation of various types of congestion management and control strategies largely depend on the measures. This study explores new possible acquisition methods of two key measures (the queue length and delay at signalized intersections). Benefited from multi-source data, an improved model for queue estimation model in HCM 2000 solves the difficulty in obtaining the initial queue length, while a shockwave-based queue estimation model rely on no high-resolution data or more than one probe vehicle data is proposed. The models are tested in field. Four commonly adopted delay estimation models are compared using field data observed from signalized intersections, the trends of the precision of these models under a variety of v/c ratios and the influential factors are studied, the applicability of these models under different situations is also analyzed.
In order to reduce the probability and risk of road traffic congestion induced by recently implemented special events, and to cope with the congestion evolution characteristics of dissemination in a certain direction gradually and significant directionality, this study first defines the concept of traffic network variable structure optimization. Then, it proposes a road section variable structure optimization method, which takes the minimum composite v/c ratio CL'C as optimization objective, to prevent traffic congestion before special events. Additionally, intersection variable structure optimization method is introduced according to road section variable structure method. Targeting at the intra-regional traffic, cross traffic, and through traffic of the influence region of special events, this study intends to mitigate traffic congestion through a group guidance strategy by assisting traffic network variable structure optimization. Also, these strategies are tested through case studies.
To deal with the accidental traffic congestion which cannot be taken into account by planning or by short time preventing strategies during special events, this study proposes a strategy system to handle area congestion, line congestion, and point congestion by implementing regional network control, path coordination control, and single point control individually. To be more specific, a boundary intersection Blocking-Discharging control strategy, which is based on fuzzy control, is proposed based on the congestion evolution characteristic of diffusion from the core to the periphery, and the fuzzy controller is designed; a path coordination control strategy, which is based on green wave and red wave concept, is proposed based on congestion evolution characteristic of dissemination in a certain direction gradually, and the determination of phase difference using graphic method is emphatically studied. These strategies are tested in simulations.
Through the understanding of traffic operation and congestion evolution characteristics during special events, and also using system science concept, this study proposes a detailed management oriented congestion management theory system. The system consists of congestion identification and assessment, which is based on evaluation measure, and congestion prevention and mitigation, which is based on temporal-spatial optimization. It enriches and improves the theories and methods of traffic management and control during special conditions. Restricted by the data, field experiment conditions, and the complexity of management theory system itself and implementation, there are some shortcomings in this study, in particular the failure to test and analyze the effects of proposed strategies in field. This might be a potential topic for further studies. At the same time, the platform development for road congestion management during special events is also an important direction in future.
Key Words: special events, congestion management, traffic network variable structure, coordination control, optimization theory
