Some results on a SIRS epidemic model with subpopulations (Q805529)

The following system of 3n autonomous ordinary differential equations has been widely used in the study of the spread of infectious diseases: \[ (1)\quad x_ i'=b_ i(N_ i-x_ i)-x_ i\sum_{j}\lambda_{ij}y_ j+\kappa_ iz_ i, \] \[ y_ i'=-(\gamma_ i+b_ i)y_ i+x_ i\sum_{j}\lambda_{ij}y_ j, \] \[ z_ i'=-(b_ i+\kappa_ i)z_ i+\gamma_ iy_ i,\quad 1\leq i\leq n, \] where \(x_ i(0)\), \(y_ i(0)\) and \(z_ i(0)\geq 0\). Here \(x_ i\) (resp. \(y_ i\); resp. \(z_ i)\) denotes the number of susceptible (resp. infected; res. recovered) individuals in the i-th subpopulation. \(N_ i\) (resp. \(b_ i\); resp, \(\gamma_ i\); resp. \(\kappa_ i)\) is the total size (resp. birth and death rate; resp. recovery rate; resp. rate at which recovered individuals lose immunity) for the i-th subpopulation. \(\lambda_{ij}\) is the effective contact rate between individuals in the i-th subpopulation with individuals in the j-th subpopulation. All the parameters \(N_ i,b_ i,\gamma_ i,k_ i,\lambda_{ij}\) are assumed to be nonnegative. A major unsolved problem in mathematical epidemiology is to determine whether the endemic equilibrium of (1), i.e. an equilibrium of the form \((x^*,y^*,z^*)\) where \(x_ i^*,y_ i^*,z_ i^*>0\), is globally stable. We will show that (i) if the effective contact rates between different subpopulations are small, i.e. \(\lambda_{ij}(i\neq j)\) is small or (ii) if \(\gamma_ i/(b_ i+\kappa_ i)<1\) for all i, then the endemic equilibrium (if it exists) is globally stable.