Abstract:

In order to determine critical links of urban road network better under emergency environment, this paper combines with network’s vulnerability analysis to structure a critical links’ identification model, which covers single and multiple failure links. Firstly, getting random attack on network and using network efficiency and maximal connected subgraphs to find potential critical links; then, introducing 0-1 variables to establish a nonlinear optimization model with road’s capacity; on the basis, handled objective function and constraints by piecewise linearization and translating the model into a mixed 0-1 programming problem; finally, using branch and bound method to solve model and an example is given to verify it. The results shows that: the critical links set is not a simple integration with single critical links, whose constituent elements don’t have adjacency relationship each other on geometric topological level. But with the number of failure links increasing, there is a significant negative correlation that can be described by reverse curve between the robustness of network and its total impedance variation, which owns well-tried effect.

Key words: traffic engineering, critical links set, mixed 0-1 programming, emergency environment, branch and bound method

摘要：

为了更好地判断出突发环境下城市道路网络中的关键路段，本文结合路网脆弱性分析方法，构建了一种涵盖单条到多条路段失效的关键路段集识别模型.首先，对路网进行随机攻击，运用网络效率、最大连通子图等鲁棒性指标寻找潜在关键路段；然后，引入0-1变量，建立一个含道路通行能力约束的非线性优化模型；再以此为基础，用分段线性化手段处理目标函数与约束条件，将模型转化为一个混合0-1规划问题；最后，以分支定界法求解模型并设计算例验证.结果表明：突发环境下的路网关键路段集一般不是若干关键路段的简单集成，其构成元素在几何拓扑层面上也不具有邻接关系；而且随着失效路段数的增多，路网鲁棒性与总阻抗变化量之间会呈现出较显著的负相关关系，可用“逆向”曲线进行刻画，效果良好.

关键词: 交通工程, 关键路段集, 混合0-1规划, 突发环境, 分支定界法