Generating and solving the mean field and pair approximation equations in epidemiological models

TL;DR

This paper introduces software tools that automate the generation and simplification of mean field and pair approximation equations in epidemiological models, streamlining the modeling process.

Contribution

It provides Perl programs to automatically generate and simplify pair approximation equations, facilitating their use in epidemiological modeling.

Findings

Automated equation generation reduces manual effort.

Tools ensure algebraic consistency and positivity in numerical solutions.

Streamlines the modeling workflow for epidemiologists.

Abstract

The pair approximation is a simple, low-order method to incorporate effects of local spatial structure in epidemiological models. However, since for K state variables in a model there are K(K+1)/2 equations in the pair approximation, generating these equations, although straightforward, can become tedious. In this paper we describe two programs written in Perl to simplify this process - one to construct the equations, and the other to generate Matlab/Octave functions to numerically integrate the equations using a positivity-preserving method. A third utility program is also included which generates a Maple file that can be used within the Maple symbolic manipulation program to simplify algebraically some of the terms in the generated file describing the equations, as well as to check that the usual combination of equations sum up to zero, which is expected in cases where the total population sizes are conserved.