从图的拓扑特征角度出发,提出了一种解决交通网络线圈布局优化问题的支撑树算法。新方法首先通过网络超级变换使得流量守恒条件在变换后网络的所有节点上均成立,再通过分析一般支撑树的拓扑特征,从而得到一种解决全网观测最小线圈集位置确定问题的图论方法。新方法具有如下特点:a)避免了复杂的数学分析与运算;b)揭示了安装线圈路段的网络拓扑结构;c)利用支撑树多样性的优势,可提供多种优化布局方案。最后通过算例验证了该方法的合理性与实用性。
From a perspective of topology, this paper proposed a spanning tree approach to solve the network sensor location problems. The new approach transformed the original traffic network to a network with only one virtual centroid. The flow conservation condition held at every node in the transformed network. Through analyzing the properties of spanning tree,it presented the new method to identify sensor locations for whole link flow inference. The new method has following features: a)avoiding the complex algebraic analysis and computation ; b) uncovering the topological property of links installed sensors; c ) the variety of spanning trees of a connected graph supplies multiply location schemes. At last, it gave a numerical example to illustrate the rationality and practicability of the new method.