Pawel Pralat

Emergence of segregation in evolving social networks

In many social networks, there is a high correlation between the similarity of actors and the existence of relationships between them. This paper introduces a model of network evolution where actors are assumed to have a small aversion from being connected to others who are dissimilar to themselves, and yet no actor strictly prefers a segregated network. This model is motivated by Schelling’s [Schelling TC (1969) Models of segregation. Am Econ Rev 59:488–493] classic model of residential segregation, and we show that Schelling’s results also apply to the structure of networks; namely, segregated networks always emerge regardless of the level of aversion. In addition, we prove analytically that attribute similarity among connected network actors always reaches a stationary distribution, and this distribution is independent of network topology and the level of aversion bias. This research provides a basis for more complex models of social interaction that are driven in part by the underlying attributes of network actors and helps advance our understanding of why dysfunctional social network structures may emerge.

A spatial web graph model with local influence regions

The web graph may be considered as embedded in a topic space, with a metric that expresses the extent to which web pages are related to each other. Using this assumption, we present a new model for the web and other complex networks, based on a spatial embedding of the nodes, called the Spatial Preferred Attachment (SPA) model. In the SPA model, nodes have influence regions of varying size, and new nodes may only link to a node if they fall within its influence region. We prove that our model gives a power law in-degree distribution, with exponent in (2, ∞) depending on the parameters, and with concentration for a wide range of in-degree values. We also show that the model allows for edges that span a large distance in the underlying space, modelling a feature often observed in real-world complex networks.

Some typical properties of the spatial preferred attachment model

We investigate a stochastic model for complex networks, based on a spatial embedding of the nodes, called the Spatial Preferred Attachment (SPA) model. In the SPA model, nodes have spheres of influence of varying size, and new nodes may only link to a node if they fall within its influence region. The spatial embedding of the nodes models the background knowledge or identity of the node, which influences its link environment. In this paper, we focus on the (directed) diameter, small separators, and the (weak) giant component of the model.

Pursuit-Evasion in Models of Complex Networks

Pursuit-evasion games, such as the game of Cops and Robbers, are a simplified model for network security. In this game, cops try to capture a robber loose on the vertices of the network. The minimum number of cops required to win on a graph G is its number. We present asymptotic results for the game of Cops and Robbers played in various stochastic network models, such as in G(n, p) with nonconstant p and in random power-law graphs. We find bounds for the cop number of G(n, p) for a large range p as a function of n. We prove that the cop number of random power-law graphs with n vertices is asymptotically almost surely Θ(n). The cop number of the core of random power-law graphs is investigated, and it is proved to be of smaller order than the order of the core.

A geometric model for on-line social networks

We study the link structure of on-line social networks (OSNs), and introduce a new model for such networks which may help infer their hidden underlying reality. In the geo-protean (GEO-P) model for OSNs nodes are identified with points in Euclidean space, and edges are stochastically generated by a mixture of the relative distance of nodes and a ranking function. With high probability, the GEO-P model generates graphs satisfying many observed properties of OSNs, such as power law degree distributions, the small world property, densification power law, and bad spectral expansion. We introduce the dimension of an OSN based on our model, and examine this new parameter using actual OSN data. We discuss how the dimension parameter of an OSN may eventually be used as a tool to group users with similar attributes using only the link structure of the network.

Meyniel's conjecture holds for random graphs

In the game of cops and robber, the cops try to capture a robber moving on the vertices of the graph. The minimum number of cops required to win on a given graph G is called the cop number of G. The biggest open conjecture in this area is the one of Meyniel, which asserts that for some absolute constant C, the cop number of every connected graph G is at most C|VG|. In this paper, we show that Meyniels conjecture holds asymptotically almost surely for the binomial random graph Gn,p, which improves upon existing results showing that asymptotically almost surely the cop number of Gn,p is Onlogn provided that pni¾?2+elogn for some e>0. We do this by first showing that the conjecture holds for a general class of graphs with some specific expansion-type properties. This will also be used in a separate paper on random d-regular graphs, where we show that the conjecture holds asymptotically almost surely when d=dni¾?3.

Geometric graph properties of the spatial preferred attachment model

The spatial preferred attachment (SPA) model is a model for networked information spaces such as domains of the World Wide Web, citation graphs, and on-line social networks. It uses a metric space to model the hidden attributes of the vertices. Thus, vertices are elements of a metric space, and link formation depends on the metric distance between vertices. We show, through theoretical analysis and simulation, that for graphs formed according to the SPA model it is possible to infer the metric distance between vertices from the link structure of the graph. Precisely, the estimate is based on the number of common neighbours of a pair of vertices, a measure known as co-citation. To be able to calculate this estimate, we derive a precise relation between the number of common neighbours and metric distance. We also analyse the distribution of edge lengths, where the length of an edge is the metric distance between its end points. We show that this distribution has three different regimes, and that the tail of this distribution follows a power law.

Geometric Protean Graphs

Abstract We study the link structure of online social networks (OSNs) and introduce a new model for such networks that may help in inferring their hidden underlying reality. In the geo-protean (GEO-P) model for OSNs, nodes are identified with points in Euclidean space, and edges are stochastically generated by a mixture of the relative distance of nodes and a ranking function. With high probability, the GEO-P model generates graphs satisfying many observed properties of OSNs, such as power-law degree distributions, the small-world property, densification power law, and bad spectral expansion. We introduce the dimension of an OSN based on our model and examine this new parameter using actual OSN data. We discuss how the geo-protean model may eventually be used as a tool to group users with similar attributes using only the link structure of the network.

Search Algorithms for Unstructured Peer-to-Peer Networks

We study the performance of several search algorithms on unstructured peer-to-peer networks, both using classic search algorithms such as flooding and random walk, as well as a new hybrid algorithm proposed in this paper. This hybrid algorithm first uses flooding to find sufficient number of nodes and then starts random walks from these nodes. We compare the performance of the search algorithms on several graphs corresponding to common topologies proposed for peer- to-peer networks. In particular, we consider binomial random graphs, regular random graphs, power-law graphs, and clustered topologies. Our experiments show that for binomial random graphs and regular random graphs all algorithms have similar performance. For power-law graphs, flooding is effective for small number of messages, but for large number of messages our hybrid algorithm outperforms it. Flooding is ineffective for clustered topologies in which random walk is the best algorithm. For these topologies, our hybrid algorithm provides a compromise between flooding and random walk. We also compare the proposed hybrid algorithm with the fc-walker algorithm on power-law and clustered topologies. Our experiments show that while they have close performance on clustered topologies, the hybrid algorithm has much better performance on power-law graphs. We theoretically prove that flooding is effective for regular random graphs which is consistent with our experimental results.

Growing Protean Graphs

The web may be viewed as a graph each of whose vertices corresponds to a static HTML web page and each of whose edges corresponds to a hyperlink from one web page to another. Recently, there has been considerable interest in using random graphs to model complex real-world networks to gain an insight into their properties. In this paper we propose an extended version of a new random model of the web graph in which the degree of a vertex depends on its age. We use the differential equation method to obtain basic results on the probability of edges being present. From this we are able to characterize the degree sequence of the model and study its behaviour near the connectivity threshold.