Figure 3
The top graph is titled “Single-node disruption”. The horizontal axis is labeled in percentages and ranges from 0 percent to 100 percent in increments of 10 percent. The vertical axis lists scenarios labeled “M S O M 18” through “M S O M 27”. Data points are shown for six strategies as indicated by the legend: Random, Node Degree, Degree Centrality, Closeness Centrality, Betweenness Centrality, and Page Rank. Random values cluster at the lower end of the scale, mostly between about 0 percent and 30 percent, with several points near 10 percent to 20 percent. Node Degree, Degree Centrality, and Closeness Centrality form mid-range clusters primarily between about 35 percent and 60 percent across the M S O M cases, with some values extending slightly above 60 percent. Betweenness Centrality and Page Rank appear at the higher end of the distribution, with most values ranging from about 65 percent to above 90 percent, and several points concentrated between roughly 90 percent and 95 percent. The bottom graph is titled “Multi-node disruption”. The horizontal axis ranges from 0 percent to 50 percent in increments of 5 percent, and the vertical axis again lists “M S O M 18” through “M S O M 27”. The same six strategies are shown. Random values cluster at the lowest range, generally between about 10 percent and 20 percent. Node Degree, Degree Centrality, and Closeness Centrality form middle clusters mostly between about 18 percent and 28 percent. Betweenness Centrality and Page Rank appear at the higher end relative to other strategies, typically between about 25 percent and 45 percent, with a noticeably narrower spread than in the single-node disruption plot. Note: All numerical values are approximated.Average reduction of performance loss per node fortification strategy on different MSOM networks