Introduction to the Modeling and Analysis of Complex Systems

352 CHAPTER 16. DYNAMICAL NETWORKS I: MODELINGg.pos = nx.spring_layout(g)g.count = 0def observe():global gcla()nx.draw(g, pos = g.pos)def update():global gg.count += 1if g.count % 20 == 0: # rewiring once in every 20 stepsnds = g.nodes()i = rd.choice(nds)if g.degree(i) > 0:g.remove_edge(i, rd.choice(g.neighbors(i)))nds.remove(i)for j in g.neighbors(i):nds.remove(j)g.add_edge(i, rd.choice(nds))# simulation of node movementg.pos = nx.spring_layout(g, pos = g.pos, iterations = 5)import pycxsimulatorpycxsimulator.GUI().start(func=[initialize, observe, update])The initialize function creates a ring-shaped network topology by connecting eachnode to its k nearest neighbors (j ∈ {1, . . . , k/2}). “% n” is used to confine the namesof nodes to the domain [0,n−1] with periodic boundary conditions. An additional attributeg.count is also created in order to count time steps between events of topologicalchanges. In the update function, g.count is incremented by one, and the random edgerewiring is simulated in every 20 steps as follows: First, node i is randomly chosen, andthen one of its connections is removed. Then node i itself and its neighbors are removedfrom the candidate node list. A new edge is then created between i and a node randomlychosen from the remaining candidate nodes. Finally, regardless of whether a randomedge rewiring occurred or not, node movement is simulated using the spring_layout