Increasing stability with complexity in a system composed of unstable subsystems

Dynamical systems and ergodic theory (37-XX) General biology and biomathematics (92B05) Physiological, cellular and medical topics (92Cxx) Ecology (92D40) Linear systems in control theory (93C05) Control/observation systems governed by ordinary differential equations (93C15) Asymptotic stability in control theory (93D20)

Review text

We examine stability of \textit{G. W. Hoffmann}'s symmetric model of the immune system [in C. Delisi and J. R. J. Hiernaux (eds.), Regulation of immune response dynamics (1982)] \[ (1)\quad \dot x_ i=S_ i-x_ i\sum^{n}_{j=1}K_{ij}x_ j;\quad x_ i>0;\quad i=1,2,...,n; \] where \(S_ i>0\), \(K_{ij}=K_{ji}\geq 0\). This paper gives necessary and sufficient conditions on \(\{S_ i\}\) and \(\{K_{ij}\}\) for Eq. (1) to have a unique, stable, steady-state solution. Determining existence of a steady-state solution requires a theorem delimiting the range R of a function \({\mathbb{F}}: D\subseteq {\mathbb{R}}^ n\to R\subseteq {\mathbb{R}}^ n,\) where D is a (possibly proper) subset of \({\mathbb{R}}^ n\). This theorem may be new. If off-diagonal elements \(\{K_{ij}:\) \(i\neq j\}\) are nonzero with probability C and \(0<S_{\min}\leq S_ i\leq \rho S_{\min}\), \(\rho\) a fixed integer, we let P(n,C) be the probability that Eq. (1) does not have a stable, steady-state solution. Let \[ T(n)=\rho^{-1}(\rho +1)^ 2n^{-1}\ln n. \] As \(n\to \infty\), \(C/T(n)\to r>1\) implies P(n,C)\(\to 0\). If we set \(\{K_{ii}=0\); \(i=1,2,...,n\}\), this result shows that accumulating more unstable subsystems increases the probability of stability of this system.

Recommendations

Cites work

Cited in

(2)

This page was built for publication: Increasing stability with complexity in a system composed of unstable subsystems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1073729)