15.08.2015 Views

Introduction to the Modeling and Analysis of Complex Systems

introduction-to-the-modeling-and-analysis-of-complex-systems-sayama-pdf

introduction-to-the-modeling-and-analysis-of-complex-systems-sayama-pdf

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

412CHAPTER 18. DYNAMICAL NETWORKS III: ANALYSIS OF NETWORK DYNAMICSNote that, although α <strong>and</strong> λ i are both non-negative, H ′ (x s (t)) could be ei<strong>the</strong>r positive ornegative, so which inequality is more important depends on <strong>the</strong> nature <strong>of</strong> <strong>the</strong> output functionH <strong>and</strong> <strong>the</strong> trajec<strong>to</strong>ry x s (t) (which is determined by <strong>the</strong> reaction term R). If H ′ (x s (t))always stays non-negative, <strong>the</strong>n <strong>the</strong> first inequality is sufficient (since <strong>the</strong> second inequalitynaturally follows as λ 2 ≤ λ n ), <strong>and</strong> thus <strong>the</strong> spectral gap is <strong>the</strong> only relevant information<strong>to</strong> determine <strong>the</strong> synchronizability <strong>of</strong> <strong>the</strong> network. But if not, we need <strong>to</strong> consider both <strong>the</strong>spectral gap <strong>and</strong> <strong>the</strong> largest eigenvalue <strong>of</strong> <strong>the</strong> Laplacian matrix.Here is a simple example. Assume that a bunch <strong>of</strong> nodes are oscillating in an exponentiallyaccelerating pace:dθ idt = βθ i + α ∑ j∈N i(θ j − θ i ) (18.16)Here, θ i is <strong>the</strong> phase <strong>of</strong> node i, <strong>and</strong> β is <strong>the</strong> rate <strong>of</strong> exponential acceleration that homogeneouslyapplies <strong>to</strong> all nodes. We also assume that <strong>the</strong> actual values <strong>of</strong> θ i diffuse <strong>to</strong> <strong>and</strong>from neighbor nodes through edges. Therefore, R(θ) = βθ <strong>and</strong> H(θ) = θ in this model.We can analyze <strong>the</strong> synchronizability <strong>of</strong> this model as follows. Since H ′ (θ) = 1 > 0,we immediately know that <strong>the</strong> inequality (18.14) is <strong>the</strong> only requirement in this case. Also,R ′ (θ) = β, so <strong>the</strong> condition for synchronization is given byαλ 2 > β, or λ 2 > β α . (18.17)Very easy. Let’s check this analytical result with numerical simulations on <strong>the</strong> Karate Clubgraph. We know that its spectral gap is 0.4685, so if β/α is below (or above) this value,<strong>the</strong> synchronization should (or should not) occur. Here is <strong>the</strong> code for such simulations:Code 18.2: net-sync-analysis.pyimport matplotlibmatplotlib.use(’TkAgg’)from pylab import *import networkx as nxdef initialize():global g, nextgg = nx.karate_club_graph()g.pos = nx.spring_layout(g)for i in g.nodes_iter():g.node[i][’<strong>the</strong>ta’] = r<strong>and</strong>om()

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!