Skip to main content
. 2019 Jul 23;9:10692. doi: 10.1038/s41598-019-47119-2

Figure 1.

Figure 1

Robustness of seven real-world complex weighted networks under different strategies of nodes-links removals. The robustness of real-world complex weighted networks was analyzed under different nodes-links removals strategies and using different measures for the system functioning. (a) Efficiency of the system (Eff) as a function of the fraction of nodes or links removed (q). The network efficiency (Eff) decreases under nodes-links removal; in this example the red strategy produced a sharper decrease in the network efficiency meaning that is more harmful than the black strategy. (b) Example of the complex networks robustness (R) for the two strategies of nodes or links removal of the outcomes in (a). The robustness (R) of the removal strategy is the area below the curve produced by the removal strategy in (a). The robustness (R) produced by the removal strategy is normalized on the max value of the strategy robustness for that system functioning measurement; in this way we can easily compare the robustness of the network under different nodes-links removal strategies. (c,d) The real-world networks robustness (R) under different nodes-links removal strategies. (c) Links removal strategies. (d) Nodes removal strategies. Eff indicates the weighted system efficiency computed on the real-world networks; Effran is the weighted efficiency computed on the real-world network after the randomization of the link weights, i.e. the real link weights are randomly re-assigned on the network links. LCC indicates the largest connected cluster in the network; LCCran is the largest connected cluster measurements computed on the real-world network after the randomization of the link weights; TF is the total flow in the network, i.e. the sum of the all link weights; TFran is the total flow on the real-world networks after the randomization of link weights. The random removal (Rand) and the randomized link weights network counterparts (LCCran, EFFran and TFran) outcomes are the average of 104 simulations.