Shunkun Yang, Qian Yang, Xiaoyun Xu, Dan Lu, Daqing Li
School of Reliability and Systems Engineering, Beihang University, Beijing, China.
Science and Technology on Reliability and Environmental Engineering Laboratory, Beijing, China.
PLoS One. 2016 Jun 14;11(6):e0156764. doi: 10.1371/journal.pone.0156764. eCollection 2016.
A network is considered observable if its current state can be determined in finite time from knowledge of the observed states. The observability transitions in networks based on random or degree-correlated sensor placement have recently been studied. However, these placement strategies are predominantly based on local information regarding the network. In this paper, to understand the phase transition process of network observability, we analyze the network observability transition for a betweenness-based sensor placement strategy, in which sensors are placed on nodes according to their betweenness. Using numerical simulations, we compute the size of the network's largest observable component (LOC) and compare the observability transitions for different sensor placements. We find that betweenness-based sensor placement can generate a larger LOC in the observability transition than the random or degree-based placement strategy in both model and real networks. This finding may help to understand the relationship between network observability and the topological properties of the network.
如果一个网络的当前状态可以根据观测状态的知识在有限时间内确定，那么该网络就被认为是可观测的。最近对基于随机或度相关传感器放置的网络中的可观测性转变进行了研究。然而，这些放置策略主要基于关于网络的局部信息。在本文中，为了理解网络可观测性的相变过程，我们分析了一种基于介数的传感器放置策略的网络可观测性转变，在这种策略中，传感器根据节点的介数放置在节点上。通过数值模拟，我们计算了网络最大可观测组件（LOC）的大小，并比较了不同传感器放置的可观测性转变。我们发现，在模型网络和真实网络中，基于介数的传感器放置在可观测性转变中都能比随机或基于度的放置策略产生更大的LOC。这一发现可能有助于理解网络可观测性与网络拓扑特性之间的关系。
