چكيده به لاتين
The idea of a route forward purchase is an innovative approach in order to control congestion in over-saturated highway networks. This approach, with a new vision in macro pricing policies, seeks to provide users with highway facilities, like other transportation facilities (air, rail, etc.) following previous policies and circumstances in an informative manner. This policy, which is recommended for peak periods (especially morning peak), requires support from the public transport sector. Therefore, this study attempts to answer questions such as 1. What is the purpose of executing a route forward purchase plan and which indicators under what conditions will perform well? 2. What would be the situation of modes meanwhile such a strict policy? What assumptions does informative assignment require? In order to present an informative assignment algorithm, machine learning methods have been used in predicting mode choices. Also, the assignment model is static, and it is implemented incrementall assignment by searching for the shortest path using the Dijkstra algorithm. The studied network is the US Sioux-Falls Network with a hypothetical public transport network consisting of three metro lines and three BRT lines. The results show that the performance of the logistic regression (LR) as a machine learning algorithm on the available modal split dataset is more fit than other algorithms. In general, it can be concluded that indicators based on revenue from a pricing plan (such as maximizing network revenue) result in significant shifts from private mode to public mode. Consequently, using such indices is recommended in networks with a high ratio of saturated links to total links. In networks with low over-saturated links, it is more appropriate to use travel time-based indices. Because indicators based on minimizing network travel time often provide less mode shift. However, using the average travel time index for different modes (〖AT〗_at) would provide more balanced answers. Often combined indices that act as multi-objective functions yield results between the two other states. That is, the shift of travel modes in multi-objective indicators such as R_t/T_t , which consider simultaneously maximizing revenue and minimizing travel time on the network, has a value between the two prior approaches.
