Abstract:Most epidemic processes on networks can be modelled by a compartmental model, that specifies the spread of a disease in a population. The corresponding compartmental graph describes how the viral state of the nodes (individuals) changes from one compartment to another. If the compartmental graph does not contain directed cycles (e.g. the famous SIR model satisfies this property), then we provide an analytic, closed-form solution of the continuous-time Markovian compartmental model on heterogeneous networks. The eigenvalues of the Markovian process are related to cut sets in the contact graph between nodes with different viral states. We illustrate our finding by analytically solving the continuous-time Markovian SI and SIR processes on heterogeneous networks. We show that analytic extensions to e.g. non-Markovian dynamics, temporal networks, simplicial contagion and more advanced compartmental models are possible. Our exact and explicit formula contains sums over all paths between two states in the SIR Markov graph, which prevents the computation of the exact solution for arbitrary large graphs.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to provide an analytical, closed - form solution for the Markov epidemic process without reinfection on heterogeneous networks. Specifically, the author focuses on the analytical solution of the continuous - time Markov compartment model on heterogeneous networks, especially when the compartment graph does not contain directed cycles, such as the property satisfied by the well - known SIR model. The main contribution of the paper lies in providing a method to analytically solve the spread of continuous - time Markov SI and SIR processes on heterogeneous networks, and showing how to extend to non - Markovian dynamics, time - varying networks, simplex contagion and more complex compartment models. In addition, the paper also discusses the computational limitations of this method, especially when dealing with arbitrarily large graphs. ### Background and Problem Description of the Paper Most epidemic processes can be modeled by compartment models, which specify how the disease spreads in a population. The corresponding compartment graph describes how the viral states of nodes (individuals) transfer from one compartment to another. If the compartment graph does not contain directed cycles (for example, the well - known SIR model satisfies this property), then an analytical, closed - form solution can be provided. ### Main Research Contents 1. **Derivation of Analytical Solutions**: - The author provides analytical solutions for the continuous - time Markov compartment model on heterogeneous networks. - The key to the analytical solution lies in that the eigenvalues of the Markov process are related to the cut sets between nodes with different viral states in the contact graph. 2. **Application Examples**: - The author demonstrates the effectiveness of the method by analytically solving the spread of continuous - time Markov SI and SIR processes on heterogeneous networks. - The physical meaning of the eigenvalues is explained in detail, and the complete solution is derived. 3. **Extensions and Limitations**: - The possibility of extending the analytical solution when dealing with non - Markovian dynamics, time - varying networks, simplex contagion and more complex compartment models is discussed. - The computational limitations of the analytical solution are pointed out, especially when dealing with arbitrarily large graphs, where it is necessary to sum over all paths, which may not be feasible in practical applications. ### Mathematical Formulas - **Calculation of Eigenvalues**: \[ \lambda_i = -\sum_{k \in I_i} \sum_{l \in S_i} \tilde{\beta}_{kl} \] where \( I_i \) represents the set of infected nodes in configuration \( i \), \( S_i \) represents the set of susceptible nodes in configuration \( i \), and \( \tilde{\beta}_{kl} \) represents the weighted infection rate between node \( k \) and node \( l \). - **Form of the Solution**: \[ s(t) = \pi + \sum_{i = 1}^{2^N - 1} c_i e^{\lambda_i t} w_i \] where \( \pi \) is the steady - state vector, \( c_i = s(0)^T v_i \), and \( v_i \) and \( w_i \) are the right and left eigenvectors respectively. ### Conclusion The paper successfully provides an analytical solution for the continuous - time Markov compartment model on heterogeneous networks and shows its applications in various epidemic models. Although there are computational limitations, this method provides a new perspective and tool for understanding epidemic spread.

Analytic solution of Markovian epidemics without re-infections on heterogeneous networks

On the Dynamics of Deterministic Epidemic Propagation over Networks

Analytic models for SIR disease spread on random spatial networks

Epidemic models on homogeneous networks : some analytical results

Analytical solution of susceptible-infected-recovered models on homogeneous networks

Discrete-time Markov chain approach to contact-based disease spreading in complex networks

Global Analysis of an Sis Model with an Infective Vector on Complex Networks

A message passing approach for general epidemic models

Stochastic Compartment Model with Mortality and Its Application to Epidemic Spreading in Complex Networks

Dynamics of Stochastic Epidemics on Heterogeneous Networks

Real-time growth rate for general stochastic SIR epidemics on unclustered networks

An Exact Relationship Between Invasion Probability and Endemic Prevalence for Markovian SIS Dynamics on Networks

Epidemic Spreading on Heterogeneous Networks with Identical Infectivity

Consequences of non-Markovian healing processes on epidemic models with recurrent infections on networks

Adaptive Susceptibility and Heterogeneity in Contagion Models on Networks

Multi-state epidemic processes on complex networks

SIS and SIR Epidemic Models under Virtual Dispersal

Epidemics on networks with heterogeneous population and stochastic infection rates

Efficient weighted-ensemble network simulations of the SIS model of epidemics

Edge-based compartmental modelling of an SIR epidemic on a dual-layer static-dynamic multiplex network with tunable clustering