Presentation is loading. Please wait.

Presentation is loading. Please wait.

Complex Networks: Complex Networks: Structures and Dynamics Changsong Zhou AGNLD, Institute für Physik Universität Potsdam.

Similar presentations


Presentation on theme: "Complex Networks: Complex Networks: Structures and Dynamics Changsong Zhou AGNLD, Institute für Physik Universität Potsdam."— Presentation transcript:

1 Complex Networks: Complex Networks: Structures and Dynamics Changsong Zhou AGNLD, Institute für Physik Universität Potsdam

2 Summary Part - I –Characterization of Complex Networks. Part - II –Dynamics on Complex Networks. Part - III –Relevance to Neurosciences.

3 1. INTRODUCTION Reductionism and complexity Brain in ``DynamicsLand´´

4 1. INTRODUCTION Connection topology Reductionism and complexity Crystal Lattices

5 1. INTRODUCTION Internet All-to-all interactionsCrystal Lattices Reductionism and complexity Connection topology

6 1. INTRODUCTION Internet All-to-all interactionsCrystal Lattices Reductionism and complexity Connection topology Diffusion Mean field ?

7 1. INTRODUCTION Technological Networks World-Wide Web Power Grid Internet

8 1. INTRODUCTION Social Networks Friendship Net Sexual Contacts Citation Networks Movie Actors Collaboration Networks

9 1. INTRODUCTION Transportation Networks Airport Networks Road Maps Local Transportation

10 1. INTRODUCTION Biological Networks Neural Networks Genetic Networks Protein interaction Ecological Webs Metabolic Networks

11 2. NETWORKS... A food web A Unified Approach towards the Connection Topology of various Complex Systems

12 2. NETWORKS... Networks Approach Basic Graphs Symmetrical Adjacency Matrix 1 2 4 56 7 8 3 A ij = 12345678 1 01100000 2 10111000 3 11000100 4 01001010 5 01010110 6 00101001 7 00011001 8 00000110

13 2. NETWORKS... Basic GraphsDiGraphs 12345678 1 01100000 2 00101000 3 10000000 4 01001000 5 00000100 6 00101000 7 00011001 8 00000110 Non-Symmetrical Adjacency Matrix 1 2 4 56 7 8 3 A ij = Networks Approach

14 2. NETWORKS... 2 8 5 0.58 Basic GraphsDiGraphsWeighted Graphs 0.4 0.7 4.5 1.8 2.8 0.6 0.4 7.4 Networks Approach

15 2. NETWORKS... Characterization Vertex degree: k(v) Basic Graphs Friendship

16 2. NETWORKS... Clustering Coeficient: C(v) Basic Graphs Characterization Friendship

17 2. NETWORKS... Clustering Coeficient: C(v) Number of existing connections: 2 Basic Graphs Characterization Friendship

18 2. NETWORKS... Clustering Coeficient: C(v) Number of existing connections: 2 Total Number of possible connections: ½·k v ·(k v -1) = ½·(4·3) = 6 Simple Graphs Characterization Friendship

19 2. NETWORKS... Clustering Coeficient: C(v) Number of existing connections: 2 Total Number of possible connections: ½·k v ·(k v -1) = ½·(4·3) = 6 C v = 2 / 6 = 0.333 Basic Graphs Characterization Friendship

20 2. NETWORKS... Clustering Coeficient: C(v) Number of existing connections: 2 Total Number of possible connections: ½·k v ·(k v -1) = ½·(4·3) = 6 C v = 2 / 6 = 0.333 Basic Graphs How well are the neighbours connected ! Characterization Friendship

21 2. NETWORKS... Distance (pathlength) Basic Graphs Characterization i j Friendship

22 2. NETWORKS... Distance (pathlength) Basic Graphs Characterization i j Friendship

23 2. NETWORKS... Distance (pathlength) Basic Graphs Characterization i j Friendship

24 2. NETWORKS... Distance Basic Graphs Characterization Friendship

25 2. NETWORKS... Distance Basic Graphs Characterization

26 2. NETWORKS... Distance Simple Graphs Characterization

27 2. NETWORKS... Distance Simple Graphs 0 Characterization

28 2. NETWORKS... Distance Simple Graphs 0 1 1 1 Characterization

29 2. NETWORKS... Distance Simple Graphs 0 1 1 1 2 2 2 Characterization

30 2. NETWORKS... Characterization Distance Simple Graphs 0 1 1 1 2 2 2 3 3 Distance: length of the shortest paths

31 2. NETWORKS... All-to-all distance matrix: Length of the shortest paths 0 1 1 1 2 2 2 3 3 123456789 1 211222334 2 121112234 3 112221323 4 212212123 5 212121123 6 221212220 7 323112212 8 332221121 9 443332212 L ij = Characterization

32 2. NETWORKS... Other measures Simple Graphs Characterization Neighbours` degree Average degree of the neighbouring nodes Betweenness (load) The number of shortest paths passing a node or an edge

33 2. NETWORKS... General Features of Real Networks Scale-free structure Power-law distribution of degrees

34 2. NETWORKS... General Features of Real Networks Small world structure Small distance High clustering

35 2. NETWORKS... ERDOS - RÉNYI MODELL (E-R) Random Network Models Connecting a pair of nodes with probability p

36 2. NETWORKS... Degree distribution: Poissonian! ERDOS - RÉNYI MODELL (E-R) Random Network Models Mean degree K=NP

37 2. NETWORKS... Degree distribution: Giant Component: Poissonian! ERDOS - RÉNYI MODELL (E-R) Random Network Models

38 2. NETWORKS... WATTS - STROGATZ MODELL (W-S): Degree? Clustering? Pathlength? Random Network Models

39 2. NETWORKS... WATTS - STROGATZ MODELL (W-S): Random Network Models Rewiring a link with probability p

40 2. NETWORKS... WATTS - STROGATZ MODELL (W-S): Random Network Models Having shortcuts now!

41 2. NETWORKS... SMALL - WORLD NETS = –High clustering –Short distance Watts, Strogatz. Nature 393/4, 1998 WATTS - STROGATZ MODELL (W-S): Random Network Models

42 2. NETWORKS... Regular LatticeSmall-World NetRandom Graph P(k) = δ(k-Z) : Z= number of neighbours Poissonian! Comparison

43 2. NETWORKS... Comparison Average Pathlength Average Clustering

44 2. NETWORKS... EVOLVING NETWORKS, Barabási-Albert model (B-A) Ingredients: –Growing AND –Preferential attachment Random Network Models

45 2. NETWORKS... Ingredients: –Growing AND –Preferential attachment Results: –Richer-Gets-Richer – k distribution: Scale Invariant! EVOLVING NETWORKS, Barabási-Albert model (B-A) Random Network Models

46 2. NETWORKS... Barabási, Albert. Science 286 (1999) SCALE - FREE NETWORKS EVOLVING NETWORKS, Barabási-Albert model (B-A) Random Network Models

47 2. NETWORKS... Properties of the models Small-WorldLatticeScale-FreeRandom Pathlength Clustering LongShort Large Small > Large in many real scale-free networks !

48 2. NETWORKS... Small-WorldLatticeScale-FreeRandom Pathlength Clustering LongShort Large Small > Properties of the models

49 2. NETWORKS... Network Resiliance: –Highly robust agains RANDOM failure of node. Significant Impacts

50 2. NETWORKS... Network Resiliance: –Highly robust agains RANDOM failure of node. Significant Impacts

51 2. NETWORKS... Network Resiliance: –Highly robust agains RANDOM failure of node. –Highly vulnerable to deliberate attack on HUBS. Significant Impacts

52 2. NETWORKS... Network Resiliance: –Highly robust agains RANDOM failure of node. –Highly vulnerable to deliberate attack on HUBS. Significant Impacts

53 2. NETWORKS... Network Resiliance: –Highly robust agains RANDOM failure of node. –Highly vulnerable to deliberate attack on HUBS. Applications: –Inmunization in computer networks and populations Cohen et al PRL, (2000, 2002) Significant Impacts

54 2. NETWORKS... Cat cortico-cortical connections Physics collaboration network Palla et al. Nature 435, 9 (2005) Communities and Overlapping Nodes

55 3.... AND MORE REALISTIC CHARACTERIZATION Weighted and Directed Networks GraphsWeightedDirected 5.1 7.0 9.3 0.3 5.1 7.2 9.3 0.3 6.0 degree, k( ) = 4 out-degree, out-k( ) = 3 in- degree, in-k( ) = 2 intensity. S( ) = 21.7out-intensity, out-S( ) =24.9 in-intensity, in-S( ) = 12.3 Degree k Intensity In/out-degree In/out-intensity

56 3.... AND MORE REALISTIC CHARACTERIZATION Weighted Networks Are weights correlated with degrees? –NO Scientific Collaborations (SCN) –YES World-Airport-Networks (WAN)

57 3.... AND MORE REALISTIC CHARACTERIZATION Weighted Networks Are weights correlated with degrees? –NO SCN –YES WAN

58 3.... AND MORE REALISTIC CHARACTERIZATION Weighted Networks Weighted Clustering Coeficient: (WAN) Barrat et al. (2004) PNAS vol.101, 11 v v j h


Download ppt "Complex Networks: Complex Networks: Structures and Dynamics Changsong Zhou AGNLD, Institute für Physik Universität Potsdam."

Similar presentations


Ads by Google