基于随机需求的离散交通网络设计

Discrete Network Design Under Stochastic Demand

由于影响因素众多,OD需求预测存在不确定性。本研究假定OD需求矩阵元素是概率分布已知的随机变量,以双层规划理论为基础,建立基于随机需求的离散交通网络设计模型,上层模型是预算约束下的路网期望总时间最小,下层模型是每个需求情景下的用户均衡。以Monte Carlo模拟和遗传算法为计算工具,设计了随机双层组合优化问题的可行算法。Nguyen Dupuis网络的实例计算表明,考虑需求不确定性的网络设计方案与确定性条件下的结果显著不同。本研究有助于提高交通规划决策的科学性和可靠性。

In transportation planning practice the forecasting origin-destination demand is always uncertain.The OD trip matrices are taken as random variables with known probability distributions.A discrete network design model under stochastic demand is set up using bi-level programming.The upper level refers to the system planner s objective of minimizing the expected total travel times,while the flow to the upper level is obtained from the user equilibrium in the lower level for each demand realization.An algorithm based on Monte Carlo simulation and genetic algorithm is proposed to solve the stochastic bilevel combinatorial optimization problem. Numerical results on Nguyen Dupuis network show the importance of accounting demand uncertainty for making correct decision. This research is helpful to improve the robustness of transportation planning project.