Epidemic spreading and bond percolation on multilayer networks (original) (raw)

Competitive epidemic spreading over arbitrary multilayer networks

Physical Review E, 2014

This study extends the Susceptible-Infected-Susceptible (SIS) epidemic model for single-virus propagation over an arbitrary graph to an Susceptible-Infected by virus 1-Susceptible-Infected by virus 2-Susceptible (SI 1 SI 2 S) epidemic model of two exclusive, competitive viruses over a two-layer network with generic structure, where network layers represent the distinct transmission routes of the viruses. We find analytical expressions determining extinction, coexistence, and absolute dominance of the viruses after we introduce the concepts of survival threshold and absolute-dominance threshold. The main outcome of our analysis is the discovery and proof of a region for long-term coexistence of competitive viruses in nontrivial multilayer networks. We show coexistence is impossible if network layers are identical yet possible if network layers are distinct. Not only do we rigorously prove a region of coexistence, but we can quantitate it via interrelation of central nodes across the network layers. Little to no overlapping of the layers' central nodes is the key determinant of coexistence. For example, we show both analytically and numerically that positive correlation of network layers makes it difficult for a virus to survive, while in a network with negatively correlated layers, survival is easier, but total removal of the other virus is more difficult.

Temporal Percolation of the Susceptible Network in an Epidemic Spreading

PLoS ONE, 2012

In this work, we study the evolution of the susceptible individuals during the spread of an epidemic modeled by the susceptible-infected-recovered (SIR) process spreading on the top of complex networks. Using an edge-based compartmental approach and percolation tools, we find that a time-dependent quantity W S (t), namely, the probability that a given neighbor of a node is susceptible at time t, is the control parameter of a node void percolation process involving those nodes on the network not-reached by the disease. We show that there exists a critical time t c above which the giant susceptible component is destroyed. As a consequence, in order to preserve a macroscopic connected fraction of the network composed by healthy individuals which guarantee its functionality, any mitigation strategy should be implemented before this critical time t c . Our theoretical results are confirmed by extensive simulations of the SIR process.

Epidemic incidence in correlated complex networks

Physical Review E, 2003

We introduce a numerical method to solve epidemic models on the underlying topology of complex networks. The approach exploits the mean-field like rate equations describing the system and allows to work with very large system sizes, where Monte Carlo simulations are useless due to memory needs. We then study the SIR epidemiological model on assortative networks, providing numerical evidence of the absence of epidemic thresholds. Besides, the time profiles of the populations are analyzed. Finally, we stress that the present method would allow to solve arbitrary epidemic-like models provided that they can be described by mean-field rate equations. 89.75.Fb, 05.70.Jk, 05.40.a A few years ago, Watts and Strogatz [1] introduced a model able to produce networks with properties of both regular lattices and random graphs with small diameter. Their model soon led to a burst of activity in the field [2, 3], further spurred by Barabasi and collaborators who found that many seemingly diverse systems share several topological properties such as a power law behavior in their connectivity distributions when represented as networks . These complex networks are formed by a set of many elements (or nodes) that are linked together through edges (or links) if they interact directly. Empirical evidence supports that in notable networks, such as metabolic or communication webs, the probability P (k) that any node has k links to other nodes is distributed accordingly to a power law P (k) ∼ k −γ [5, 6, 7], with γ ≤ 3 in most cases.

Spreading processes in Multilayer Networks

IEEE Transactions on Network Science and Engineering, 2015

Several systems can be modeled as sets of interconnected networks or networks with multiple types of connections, here generally called multilayer networks. Spreading processes such as information propagation among users of an online social networks, or the diffusion of pathogens among individuals through their contact network, are fundamental phenomena occurring in these networks. However, while information diffusion in single networks has received considerable attention from various disciplines for over a decade, spreading processes in multilayer networks is still a young research area presenting many challenging research issues. In this paper we review the main models, results and applications of multilayer spreading processes and discuss some promising research directions.

Second look at the spread of epidemics on networks

Physical Review E, 2007

In an important paper, M.E.J. Newman claimed that a general network-based stochastic Susceptible-Infectious-Removed (SIR) epidemic model is isomorphic to a bond percolation model, where the bonds are the edges of the contact network and the bond occupation probability is equal to the marginal probability of transmission from an infected node to a susceptible neighbor. In this paper, we show that this isomorphism is incorrect and define a semi-directed random network we call the epidemic percolation network that is exactly isomorphic to the SIR epidemic model in any finite population. In the limit of a large population, (i) the distribution of (self-limited) outbreak sizes is identical to the size distribution of (small) out-components, (ii) the epidemic threshold corresponds to the phase transition where a giant strongly-connected component appears, (iii) the probability of a large epidemic is equal to the probability that an initial infection occurs in the giant in-component, and (iv) the relative final size of an epidemic is equal to the proportion of the network contained in the giant out-component. For the SIR model considered by Newman, we show that the epidemic percolation network predicts the same mean outbreak size below the epidemic threshold, the same epidemic threshold, and the same final size of an epidemic as the bond percolation model. However, the bond percolation model fails to predict the correct outbreak size distribution and probability of an epidemic when there is a nondegenerate infectious period distribution. We confirm our findings by comparing predictions from percolation networks and bond percolation models to the results of simulations. In an appendix, we show that an isomorphism to an epidemic percolation network can be defined for any time-homogeneous stochastic SIR model.

Multidimensional epidemic thresholds in diffusion processes over interdependent networks

Chaos, Solitons & Fractals, 2015

Several systems can be modeled as sets of interdependent networks where each network contains distinct nodes. Diffusion processes like the spreading of a disease or the propagation of information constitute fundamental phenomena occurring over such coupled networks. In this paper we propose a new concept of multidimensional epidemic threshold characterizing diffusion processes over interdependent networks, allowing different diffusion rates on the different networks and arbitrary degree distributions. We analytically derive and numerically illustrate the conditions for multilayer epidemics, i.e., the appearance of a giant connected component spanning all the networks. Furthermore, we study the evolution of infection density and diffusion dynamics with extensive simulation experiments on synthetic and real networks.

Shift of percolation thresholds for epidemic spread between static and dynamic small-world networks

The European Physical Journal B, 2011

The study compares the epidemic spread on static and dynamic small-world networks. They are constructed as a 2-dimensional Newman and Watts model (500 × 500 square lattice with additional shortcuts), where the dynamics involves rewiring shortcuts in every time step of the epidemic spread. We assume susceptible-infectious-removed (SIR) model of the disease. We study the behaviour of the epidemic over the range of shortcut probability per underlying bond φ = 0-0.5. We calculate percolation thresholds for the epidemic outbreak, for which numerical results are checked against an approximate analytical model. We find a significant lowering of percolation thresholds on the dynamic network in the parameter range given. The result shows the behaviour of the epidemic on dynamic network is that of a static small world with the number of shortcuts increased by 20.7 ± 1.4%, while the overall qualitative behaviour stays the same. We derive corrections to the analytical model which account for the effect. For both dynamic and static small worlds we observe suppression of the average epidemic size dependence on network size in comparison with the finite-size scaling known for regular lattice. We also study the effect of dynamics for several rewiring rates relative to infectious period of the disease.

SIS Epidemics in Multilayer-based Temporal Networks

arXiv (Cornell University), 2018

To improve the accuracy of network-based SIS models we introduce and study a multilayer representation of a time-dependent network. In particular, we assume that individuals have their long-term (permanent) contacts that are always present, identifying in this way the first network layer. A second network layer also exists, where the same set of nodes can be connected by occasional links, created with a given probability. While links of the first layer are permanent, a link of the second layer is only activated with some probability and under the condition that the two nodes, connected by this link, are simultaneously participating to the temporary link. We develop a model for the SIS epidemic on this time-dependent network, analyze equilibrium and stability of the corresponding mean-field equations, and shed some light on the role of the temporal layer on the spreading process.

Epidemics spreading in interconnected complex networks

Physics Letters A, 2012

We study epidemic spreading in two interconnected complex networks. It is found that in our model the epidemic threshold is always lower than that in any of the two standalone networks. Detailed theoretical analysis is proposed which allows quick and accurate calculations of epidemic threshold and average outbreak/epidemic size. Theoretical analysis and simulation results show that, generally speaking, the epidemic size is not significantly affected by the inter-network correlation. In interdependent networks which can be viewed as a special case of interconnected networks, however, impacts of inter-network correlation on the epidemic threshold and outbreak size are more significant.

Effect of the Interconnected Network Structure on the Epidemic Threshold

Most real-world networks are not isolated. In order to function fully, they are interconnected with other networks, and this interconnection influences their dynamic processes. For example, when the spread of a disease involves two species, the dynamics of the spread within each species (the contact network) differs from that of the spread between the two species (the interconnected network). We model two generic interconnected networks using two adjacency matrices, A and B, in which A is a 2N × 2N matrix that depicts the connectivity within each of two networks of size N , and B a 2N × 2N matrix that depicts the interconnections between the two. Using an N-intertwined mean-field approximation, we determine that a critical susceptible-infected-susceptible (SIS) epidemic threshold in two interconnected networks is 1/λ 1 (A + αB), where the infection rate is β within each of the two individual networks and αβ in the interconnected links between the two networks and λ 1 (A + αB) is the largest eigenvalue of the matrix A + αB. In order to determine how the epidemic threshold is dependent upon the structure of interconnected networks, we analytically derive λ 1 (A + αB) using a perturbation approximation for small and large α, the lower and upper bound for any α as a function of the adjacency matrix of the two individual networks, and the interconnections between the two and their largest eigenvalues and eigenvectors. We verify these approximation and boundary values for λ 1 (A + αB) using numerical simulations, and determine how component network features affect λ 1 (A + αB). We note that, given two isolated networks G 1 and G 2 with principal eigenvectors x and y, respectively, λ 1 (A + αB) tends to be higher when nodes i and j with a higher eigenvector component product x i y j are interconnected. This finding suggests essential insights into ways of designing interconnected networks to be robust against epidemics.