Symmetries and normalization in 3-compartment epidemic models I: The replacement number dynamics

TL;DR

This paper extends the unification of epidemic models by demonstrating how a complex 14-parameter model can be normalized through symmetry, revealing fundamental relations and simplifying analysis of epidemic dynamics.

Contribution

It introduces a symmetry-based normalization method for a 14-parameter epidemic model, reducing complexity and unifying various models in the literature.

Findings

Normalization reduces model parameters from 14 to 5.

Symmetry group G classifies equivalent models.

Reveals relations between different epidemic models.

Abstract

As shown recently by the author, constant population SI(R)S models map to Hethcote's classic endemic model originally proposed in 1973. This unifies a whole class of models with up to 10 parameters, all being isomorphic to a simple 2-parameter master model for endemic bifurcation. In this work this procedure is extended to a 14-parameter SSISS Model, including social behavior parameters, a (diminished) susceptibility of the R-compartment and unbalanced constant per capita birth and death rates, thus covering many prominent models in the literature. Under mild conditions, in the dynamics for fractional variables in this model all vital parameters become redundant at the cost of possibly negative incidence rates. There is a symmetry group G acting on parameter space A, such that systems with G-equivalent parameters are isomorphic and map to the same normalized system. Using (Xrep,I) as canonical coordinates, Xrep the replacement number, normalization reduces to parameter space A/G with 5 parameters only. This approach reveals unexpected relations between various models in the literature. Part two of this work will analyze equilibria, stability and backward bifurcation and part three will further reduce the number of essential parameters from 5 to 3.