基于仿真优化的超大城市内涝场景交通疏导优先级评估

Simulation-based Optimization for Priority Assessment of Traffic Dredging in Megacity Waterlogging Sections

摘要:

针对现有研究对动态交通系统刻画精度低、忽视内涝点位间相互作用等不足，提出一种基于仿真优化的城市内涝疏导优先级评估方法。以最大化路网用户平均行程时间为目标，以内涝路段为决策变量构建0?1规划问题；将疏导优先级评估转化为离散优化问题，通过搜索疏导点位组合刻画点位间相互作用。采用中观交通仿真软件DynusT进行路网精细化建模，并利用快速机器学习模型作为代理，实现仿真优化闭环算法，求解路段疏导优先级。最后，以广州市中心城区为例构建仿真模型并进行验证。结果表明，按优先级排序治理后，路网用户平均行程速度较现有方案提升28.72%，验证了方法的准确性。

Abstract:

Addressing the limitations of existing research， such as low precision in depicting dynamic traffic systems and insufficient consideration of interactions between flood points， this study presents a simulation-based optimization method for assessing urban flood mitigation priorities. By maximizing the average travel time of road network users as the objective and using flood-affected road sections as decision variables， a 0-1 programming problem is formulated to transform the priority assessment into a discrete optimization problem. The interactions between flood points are captured through the search for optimal mitigation point combinations. A mesoscopic traffic simulation tool， DynusT， is employed for detailed road network modeling， and a fast machine learning model is utilized as a surrogate to enable a closed-loop simulation optimization algorithm for determining mitigation priorities. The method is validated through a case study in the central urban area of Guangzhou. Results demonstrate that implementing mitigation measures based on the priority ranking improve the average travel speed by 28.72% compared to existing solutions， confirming the accuracy of the method .