Investigations of Agents, Transportation and Data Communication Networks

车辆交通拥堵影响到日常生活的方方面面。它会造成时间浪费、经济损失、燃油消耗增加和污染。由于资金、空间和环境的限制，扩建路网并不是解决拥堵、可靠性和安全性问题的可行方案。人们希望在信息和通信技术领域取得进展的基础上设计智能交通系统(ITSS)，例如车载自组织网络(VANETs)，原则上可以允许车辆、红绿灯或路边单元交换信息并协调它们的行为。因此，车与车之间的通信和应用先进的信息技术可以帮助优化交通流量，提高运输效率。然而，在以最优的方式在现实生活中实施其中的许多改进之前，必须通过模拟模型研究和分析它们的情景。这项研究开发了基于多智能体的仿真模型，用于在改变道路/街道网络之前评估设计方案，并研究这些改变对交通流量、拥堵和旅行时间的影响。本研究探讨如何应用资讯和通讯技术来进行车辆交通流量的监测、侦测、预测和控制。交通异常和拥堵的存在并不是车辆交通网络独有的问题。此外，它们每天都会在数据通信网络中遇到。本研究开发了数据通信网络，即分组交换网络(PSN)的仿真模型，以研究正常和异常流量条件下(例如，在分布式拒绝服务攻击下)的时空分组流量动态。它研究：(1)功能、结构和可扩展性如何影响PSN的集体动态和性能；(2)聚合网络性能指标如何很好地捕捉由异常情况引起的PSN流量和拥塞动态的变化。此外，本研究的创新之处在于：(1)基于对车载交通和数据通信网络中时空大流量模式的多尺度分析的异常交通检测方法；(2)基于对大流量时空动态的理解以及车载交通和数据通信网络中的关联和协作的流量和拥塞控制策略；(3)用于车载交通网络交通流量管理的实时可实现的流量和拥塞控制策略、拥塞检测和预测方法。讨论的研究涉及方法论和应用领域。在方法论领域，它还研究认知主体，强调最小实体，无论是在存储方面还是在逻辑原语方面，都不能解析地表达概念，也不能使用清晰的值。所描述的研究将对方法论的进步、对流动和拥堵现象的理论理解产生新的重要结果，并有助于实际应用来解决车辆运输和数据通信网络中的一些具有经济价值的实际问题。

Vehicular traffic congestion affects many aspects of everyday life. It causes waste of time, economic losses, increased fuel consumption & pollution. Due to financial, spatial and environmental constraints extending the road network is not feasible solution to solve congestion, reliability and safety problems. The hope is in designing Intelligent Transportation Systems (ITSs) build on progress in the area of information & communication technology such as Vehicular Ad Hoc Networks (VANETs) which in principle, could allow vehicles, traffic lights, or road side units to exchange information and to coordinate their behavior. Thus, vehicle-to-vehicle communication and applications of advances in information technologies could help to optimize traffic flow and make transportation more efficient. However, before many of these advances can be implemented in real life in the most optimal way their scenarios must be studied and analyzed via simulation models. This research develops multi-agent based simulation models to be used for evaluating design options prior to making changes in road/street networks and to study the effects of these changes on traffic flow and congestion, and travel times. This research investigates how application of information and communication technologies can be used for vehicular traffic flow monitoring, detection, forecasting and control. The problems of existence of anomalous traffic and congestion are not unique to vehicular transportation networks. Also, they are encountered in data communication networks everyday. This research develops simulation models of data communication networks, i.e. of Packet Switching Networks (PSNs) to study spatio-temporal packet traffic dynamics under normal and anomalous traffic conditions (e.g., under distributed denial of service attacks). It investigates: (1) how function, structure, and scalability affect collective dynamics and performance of PSNs; (2) how well aggregate network performance indicators capture changes in PSNs flow and congestion dynamics caused by anomalous conditions. Additionally, this research develops novel: (1) detection methods of anomalous traffic based on multiscale analysis of spatio-temporal large scale flow patterns in vehicular transportation and data communication networks; (2) flow and congestion control strategies based on the understanding of spatio-temporal dynamics of large scale flow patterns and on correlation and cooperation in vehicular transportation and data communication networks; (3) real-time implementable flow and congestion control strategies, congestion detection and forecasting methods for traffic flow management in vehicular transportation networks. The discussed research covers methodology and application domains. In the methodology domain it studies also cognitive agents with emphasis on minimal entities, both in terms of storage and in terms of logical primitives and unable to express concepts analytically and unable to use crisp values. The described research will produce novel results of importance to the advancement of methodology, theoretical understanding of flow and congestion phenomena and useful for practical applications to solve some real life problems of economical value in vehicular transportation and data communication networks.