Using asymptotics for efficient stability determination in epidemiological models

Abstract

Analytical stability calculation is done to prove stability properties for systems with parameters that do not have explicit values. For systems with three components, the usual method of finding the characteristic polynomial as the determinant of J-lambda I and applying the Routh-Hurwitz conditions is reasonably efficient. For larger systems of four to six components, the method is impractical, as the calculations become too messy. In epidemiological models, there is often a very small parameter that appears as the ratio of a disease-based time scale to a demographic time scale; this allows efficient use of asymptotic approximation to simplify the calculations at little cost. Here we describe the tools and an example of efficient stability analysis, followed by a set of guidelines that are generally useful in applying the method.