In many complex networks, the main task is to transfer load from sources to destinations. Therefore, the network throughput is a very important indicator to measure the network performance. In order to improve the network throughput, researchers have proposed many effective routing strategies. Spatial networks, as a class of complex networks, exist widely in the real-world. However, the existing routing strategies in complex networks cannot achieve good results when applied in spatial networks. Therefore, in this paper, we propose a new degree-location ([Formula: see text]) routing strategy to improve the throughput of spatial networks. In this routing strategy, the load transmitted from sources to destinations will bypass the nodes with high degrees and the nodes located close to the center of region. Simulations on homogeneous and heterogeneous spatial networks show that the [Formula: see text] routing strategy proposed in this paper can effectively improve the throughput of the network. The result of this paper can help the routing design of spatial networks and may find applications in many real-world spatial networks to improve the transmission performance.

No abstract

Many networks in the real-world have spatial attributes, such as the location of nodes and the length of edges, called spatial networks. When these networks are subject to some random or deliberate attacks, some nodes in the network fail, which causes a decline in the network performance. In order to make the network run normally, some of the failed nodes must be recovered. In the case of limited recovery resources, an effective key node identification method can find the key recovering node in the failed nodes, by recovering which the network performance can be recovered the most. In this paper, we propose two key recovering node identification methods for spatial networks, which are the Euclidean-distance (ED) recovery method and route-length (RL) recovery method, respectively. Simulations on homogeneous and heterogeneous spatial networks show that the proposed methods can significantly recover the network performance.