Abstract: According to the characteristic of limited link capacity and demand capacity of travel destination, the paper proposes a Logit stochastic user equilibrium fixed model with dual constraints. The fixed model is solved by Lagrangian multiplier algorithm which can converge rapidly by mean of regulating the Lagrangian multiplier reasonably. In the iteration process, an improved method of successive weighted averages with self-regulated averaging schemes is designed to solve the general Logit stochastic equilibrium problem and can ensure gradually the links flow lower than corresponding capacity. The method can also improve the service efficiency by avoiding the route numeration. The model and algorithm are then validated by a network example.

Key words: transportation planning, demand management, Logit stochastic user equilibrium, Lagrangian multiplier algorithm

摘要： 依据现实交通网络中路段容量与出行终点停车容量空间有限性的特征，建立带路段流量和终点需求双约束的Logit随机用户均衡问题的不动点模型，设计了一种有效的Lagrangian乘子法来求解，通过合理调整Lagrangian乘子使算法快速趋于收敛。在算法的迭代过程中，对通常Logit均衡问题则设计改进的自适应相继加权平均法来求解，使路段流量不超过相应路段容量并避免了繁琐的路线枚举，改进了算法的计算效率。数值实验验证了算法的有效性和结果的可行性。

关键词: 交通规划, 需求管理, Logit随机用户均衡, Lagrangian乘子法