Abstract: Considering the differences between day-to-day traffic demand and the traffic demand of special important activities, the bi-level programming model with two lower level models for reconstructing road network is set up. The upper level model is to determine the enlarged capacities of the road network with the targets of minimizing reconstruction expense and saturation of road sections, and the lower level models are to calculate the flow of road sections under the two types of traffic demands via the user equilibrium traffic assignment method. The chaos simulated annealing algorithm is designed for solving the bi-level model. Numerical experiment demonstrates that before reconstruction, the proportion of saturated road sections is 8.3％ and 41.7％, respectively; after reconstruction that considers these two types of traffic demands, the proportion of saturated road sections declines to 0 and 8.3％; while only considering the day-to-day traffic demand, the saturated proportion is 8.3％ and 33.3％, respectively. The result shows that the reconstructed road network that considers both types of traffic demands can alleviate the traffic congestion more effectively.

Key words: urban traffic, special important activities, reconstruction of road network, bi-level programming model with two lower level models, chaos simulated annealing algorithm

摘要： 由于城市交通需求在日常和大型活动期间有较大差异，建立考虑大型活动影响下的路网改造一主多从双层规划优化模型，上层规划以城市路网改造成本和改造后路段饱和度最小为目标确定路段改造能力，下层规划分别考虑日常和大型活动两种交通需求下的用户均衡模型，并设计了基于混沌的模拟退火求解算法. 算例分析中，路网改造前的饱和路段比率分别为8.3％和41.7％；同时考虑日常和大型活动两种交通需求改造后，饱和路段比率分别为0和8.3％；单独考虑日常交通需求改造后，饱和路段比率分别为8.3％和33.3％. 这表明同时考虑两种需求改造后的路网能更有效地缓解交通拥堵.

关键词: 城市交通, 大型活动, 路网改造, 一主多从双层规划, 混沌模拟退火算法