Abstract: According to travelers' risk aversive route choice behaviors under a stochastic road network environment, a second-best congestion pricing bi-level programming model considering both network reliability and spatial equity is proposed. In the upper level model, the optimization objective is to maximize the social welfare of the road network with the spatial equity constraints, while the lower level model is an elastic demand user equilibrium model with travel time reliability taken into account in the presence of congestion pricing. In consideration of the complexity of bi-level programming model, a combined algorithm based on genetic algorithm and Frank-Wolfe algorithm is presented to solve the proposed model. The results of numerical experiments show that, by comparison with traditional second-best congestion pricing model, the proposed model with travel time reliability taken into account can generate evidently different equilibrium flow distribution patterns, which reflects that travelers' route choice behaviors could directly influence the results of congestion pricing. In addition, the results suggest that genetic algorithm is quite robust to different parameter settings.

Key words: traffic economy, second-best congestion pricing, bi-level programming, road traffic, genetic algorithm, network reliability, spatial equity

摘要： 根据随机路网环境下出行者规避风险的路径选择行为,提出了一种考虑路网可靠性和空间公平性的次优拥挤收费双层规划模型。其中,上层模型以具有空间公平性约束条件下最大化路网的社会福利为目标,下层模型是实施拥挤收费条件下考虑行程时间可靠性的弹性需求用户平衡模型。鉴于双层规划模型的复杂性,设计了基于遗传算法和FrankWolfe算法的组合式算法来求解提出的模型。算例结果表明:考虑行程时间可靠性的次优拥挤收费会产生不同于传统次优拥挤收费的平衡流量分布模式,表明出行者的路径选择行为对拥挤收费结果会产生直接影响;此外,算例结果还说明遗传算法对参数设置具有很强的鲁棒性。

关键词: 交通经济, 次优拥挤收费, 双层规划, 道路交通, 遗传算法, 网络可靠性, 空间公平性