عنوان مقاله [English]
Traffic accidents decrease the capacity of transportation networks and increase the travel delays. Advanced traveler information systems (ATIS) provide the drivers with information about these accidents, so that the uncertainty about the capacity of the network eliminates and drivers can choose better routes to their destinations. Congestion pricing is among the traffic management policies, for which considering traffic flow variation is necessary for devising accurate plans. The performance of pricing policies is directly affected by the availability of the information to the drivers in the presence of the uncertainty of the network situations. In this paper, an analytical dynamic traffic assignment model is proposed which can take account of tolls charged to drivers. An iterative algorithm is also proposed to solve the model for large-scale networks. Moreover, the assignment model is used as the sub-model within an optimization model that is proposed for dynamic cordon congestion pricing problem. To solve the pricing model, a solution algorithm is developed which works based on a grid search method, and is tuned to reach the desired accuracy level in solving the problem. Two pricing strategies are considered: tolls are considered fixed for all days of the analysis period, or vary in response to the conditions of the network for each day. Also, dependent on the provision of information to the drivers, they can be either aware or unaware of the accidents in each day. The problem is solved for combinations of different states of pricing (fixed or responsive) and information (no or full) for the network of Sioux-Falls. Numerical results show that providing information to the drivers becomes more important than performing the congestion pricing when the level of uncertainty increases, and especially when severe accidents occurs. Besides, applying fixed tolls is no longer beneficial when drivers are equipped with ATIS. Applying responsive pricing and ATIS can reduce the total travel time of the network by 11.8 percent. Moreover, some results of sensitivity analysis of accident severity and accident probability are reported, showing that as accident severity increases, fixed tolls become more useful for congestion reduction.