考虑用户微观偏好的日变交通调整模型研究

Traffic congestion has become a serious problem that influences the human settlement and the sustainable development of our urban cities. Considering the deficiencies in the existing relevant studies, this project argues that it is not only the well-known macro-behavior rule (i.e., the rational behavior swapping axiom) but also the micro-preferences (e.g., the route capacity and familiarity preferences) should be considered in modeling traveler’s day-to-day route swapping behavior. Along this line, this project plans to investigate the characteristics and action mechanism of above two micro-preferences by organizing virtual route-swapping experiments, and then apply multi-discipline theories to build several intuitively reasonable day-to-day traffic swapping models which are capable of uniting the macro-behavioral rules and micro-preferences and avoiding over-swapping problem naturally. After that, these models will be calculated and examined by data acquired from experiments. Based on the built swapping models, two works are also predesigned. Firstly, two traffic swapping algorithms, which are able to avoid over-swapping problem naturally, will be developed by establishing new parameter auto-modifying separators according to different swapping rules. Secondly, we also intend to develop a new and more effective period-to-period dayparting congestion pricing strategy to regulate the day-to-day traffic evolution of a non-equilibrium network. These studies are expected to provide a more scientific theory basis to analyze and model the non-equilibrium network, and to provide new choices to solve the network traffic equilibrium problems, as well as to provide new approach and scientific decision-making basis to alleviate the congestion of non-equilibrium traffic network. Therefore, all of these studies will be of important theoretical significance and practical values as well.

交通拥堵已成为影响我国城市人居环境和可持续发展的严重问题。本项目针对现有研究不足，提出日变（day-to-day）路径调整行为建模不仅应遵守理性行为调整这一宏观行为准则，还应同时考虑出行者的路径能力偏好、熟悉度偏好等微观偏好，为此，拟通过组织虚拟交通调整实验，深入研究出行者这两种微观偏好的特性与作用机理，再综合运用多学科知识，建立宏观行为准则与微观偏好相融、直观合理并能自行规避过度调整问题的日变交通调整模型，并利用实验数据对模型进行校验；然后在此基础上：针对不同调整行为导向，构建基于参数自动调节机制并能自行规避过度调整问题的新型流量调整算法；研讨一种新的更为有效的周期性分时段动态路段收费策略。项目的完成，可为非平衡城市交通网络分析与建模提供更科学的理论基础，为交通流均衡模型的求解提供新途径，为非平衡城市交通拥堵治理提供新方法和科学的决策依据，具有重要的理论意义与潜在实践价值。

城市交通系统是一个开放系统，容易遭遇来自外界的干扰，致使网络交通流时常处于非平衡状态。城市交通网络的非平衡常态对现有的平衡分析方法提出了挑战，要求发展出一套合服现实需要的非平衡分析理论。然而，现有的非平衡分析方法仍存在诸多不足，比如涵盖关键行为范式不全、科学管理措施匮乏等。本项目针对这些问题和不足，围绕随机和确定环境下的逐日交通调整行为规律与机理建模、流量调整算法设计以及动态交通管理策略开发等问题深入开展研究，已取得了若干创新研究成果：项目提出的PdUE比传统随机用户均衡理论更符合个体行为理性，可为信息环境下的交通网络静态规划与动态管理提供更合理的理论基础；有关逐日交通调整模型的研究，明晰了个体有限理性、微观偏好、前瞻思维等重要因素对于交通选择行为的作用机制，有助于拓展和深化对于宏观逐日交通演化复杂现象及其内在规律的认识和理解，进而为非平衡交通网络分析提供支撑模型；基于流量调整机制的交通分配算法简便易行且不依赖于导数信息，可为各类交通流均衡问题的快速近似求解提供新途径；有关逐日动态交通信息诱导与周期性动态收费的决策支持模型研究可为现实城市动态交通管理决策提供理论指导和参考依据。大部分研究成果已发表，剩余工作均已投稿在审或即将结稿。项目基本实现了预期研究目标，已发表SCI检索期刊论文4篇（其中2篇同时又被SSCI收录，并已分别获得SCI他引2次（施引者包括一名AAAS Fellow）和4次）、北大中文核心期刊论文1篇、国际学术会议论文2篇（其中1篇已被EI检索），其中5篇期刊论文的第一作者均是项目负责人；围绕探索性研究内容发表SCI检索论文3篇（其中1篇获得SCI他引3次）、国际学术会议论文1篇。项目已培养硕士研究生2人；项目组成员江世雄获得2018年度博士研究生国家奖学金；项目负责人张文义晋升副教授。有关研究成果的详细列表可参见本报告的“研究成果部分”。