Identification of efficient observers for locating spreading source in complex networks

Estimating the location of the spreading source of complex networks is a challenging task and plays an important role in many real problems. The hidden source can be localized based on the information gathered by a few nodes, which are called the observers. Identification of the efficient observers is critical to locate the source with high accuracy. Here we analyze several placement strategies of the observers based on centralities of nodes, including the high-degree, high-betweenness, high-clustering coefficient, high-eigenvector and high-closeness. Based on the random spreading experiments on both model and real networks, we find that the localization accuracy of these strategies is decreased with the increase of the connectivity of network, and there is no significant difference between them. Further experiments show that the coverage range of the observers may be the key factor that affects the localization accuracy. Our results can provide a route for the optimal design of placement strategies of the observer nodes.