AbstractsEngineering

The route guidance information, which aims at assigning travellers to a set of paths from origins to destinations, can be obtained by solving the Dynamic Traffic Assignment (DTA) problems. However, route choices given by dynamic user equilibrium (DUE) or dynamic system optimum (DSO) cannot simultaneously benefit both sides and can easily result in recurring congestion. In this context, this research focuses on the real-time traffic re-distribution and route guidance system, which can achieve an efficient distribution of network capacity over time and space without seriously violating travellers' preferences. This research is involved with the integration of traffic flow dynamics model, traffic assignment model and route choice modelling. The real-time traffic information provision system is the basis of the research, which can provide real-time traffic information prediction based on day-to-day dynamics and current traffic information. Traffic dynamics models are targeted to support Advanced Traffic Management and Information System (ATMIS) to provide a mathematical approximation of the dynamics of traffic flows based on the real-time traffic information. In this research, a path based flow model integrated the point queue model (PQ) with generalized expansion method (GEM) is proposed, which can capture traffic dynamics and model dynamic flow propagation process. Based on the path based flow model, the path marginal cost can be achieved and time dependent traffic assignment can be established. In order to meet the requirements of travellers, a route choice modelling combining the analytical hierarchy process (AHP) with fuzzy inference technique is introduced, which can handle the vagueness of traffic attributes, simplify the definition of decision strategy and represent multiple criterions explicitly. The user equilibrium (UE) traffic assignment based on the route choice modelling can be favourable to travellers. For the traffic management, it requires vehicles cooperatively to achieve system optimal (SO) traffic assignment, while traffic users competitively make route choices based on UE traffic assignment. Thus, the cooperation and competition relationship can be modelled by the Stackelberg game based model to balance the benefit between traffic authorities and travellers. Moreover, the gradient projection method is introduced to improve the efficiency of the integrated system.