Abstract： A mathematical model for quantitative generation of urban rail transit planning network is established in this study. The model is developed in consideration of enhancing the competitiveness of passenger flow sharing for rail transit, targeting at maximizing the average load intensity of the network with the network topology, node and section repeatability, line conditions, passenger flow balance, transfer condition, network scale and comfort level as constraints. Based on the main passenger corridors and passenger distributing centers, a strategy for determining the search range and search direction of the network is proposed to limit the trend boundary of the alignments reasonably. The problem belongs to NP-Hard problem, which is difficult to be solved for the large-scale network due to its large search space. So, a neighborhood search algorithm is designed based on simulated annealing framework. With the X city taken as an example, the routes of the rail transit network in the long-term are calculated. Compared with the existing planned network of the city, the network generated based on the present model has smaller scale and higher utilization rate while meeting the passenger demand along the main passenger corridors and main axis of urban development, thus verifying the effectiveness of the model and the algorithm.
