4.Results for random graphs Scaling relation of Sro p, coevolution from initial random graph with the mechanism of IVM p~Bf(k/B),B≠0 (5) y=-1.27,a=0.15
4.Results for random graphs • Scaling relation of SRO , coevolution from initial random graph with the mechanism of IVM. • • (5) ~ ( / ), 0 f k = − = 1.27, 0.15
4. Results for small world networks a (b 0.78 u.54 09 n 0 k 09 481212243832 0.0620.40.608I,6 H (d) ······· ,加 RMn·· k 4·07 C=092P·5 0.10 121610242832 k/β FIG. 3:(color online)(a)stationary efficiency p (b) LRO. (c)primitive p(upper panel)and rescaled p(k/B)(lower panel). (d) global efficiency Glob. Results are averaged over 100(5 for Eglob) configurations evolved from Sw with
4.Results for small world networks
4. Results for small world networks Scaling relation of SRo P, coevolution from initial small-world with the mechanism of vm P~g(k/B),B≠0 6) c=-0.92
4.Results for small world networks • (6) ~ ( / ), 0 g k Scaling relation of SRO , coevolution from initial small-world with the mechanism of IVM. = −0.92
4. Results and physical meanings a (b) 数4 c (d) 6=,05 0.5 RG FIG. 4:(a) SRO of SW at B=0, initial =0.1;(b)SRO for different initial at B=1;(c)SRO of RG at B=0;(d) Scaling of p of SW in(b). All with N= 2000 and averaged with 100 configurations
4.Results and physical meanings
4. Results and physical meanings For the case of B=l, another scaling relation is found for the coevolution from initial sw p~h(k)/q°2 k-k the critical average degree, h(x) is universal function and h is K where k, C=-05 is a constant, d 0.05 It reflects the correlation between evolution processes starting from different initial rewiring probability which related to effective dimension of swns It is impossible to see E=1 no matter coevolution starting from RG or SW when B=0
4.Results and physical meanings • For the case of ，another scaling relation is found for the coevolution from initial SW • (7) • where ， c= -0.5 is a constant, ，and is the critical average degree, h(x) is universal function. • It reflects the correlation between evolution processes starting from different initial rewiring probability which related to effective dimension of SWNs. It is impossible to see E=1 no matter coevolution starting from RG or SW when . c k =1 ~ ( ) / c h k c c k k − k = = −0.05 = 0