Investigations of Agents, Transportation and Data Communication Networks

点击翻译按钮获取中文摘要

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.