The Floquet exponent for two-dimensional linear systems with bounded coefficients (Q1092310)

The purpose of the paper is to define and study the Floquet exponent for the two-dimensional linear differential equations with bounded coefficients \[ (1)\quad u'=\left[ \begin{matrix} \bar a(t)\\ -\bar b(t)+\bar c(t)\end{matrix} \begin{matrix} \bar b(t)+\bar c(t)\\ -\bar a(t)\end{matrix} \right]u\equiv \bar A(t)u. \] It is defined by using ergodic theory. For this purpose the equation (1) is embedded in the family of differential equations \[ (1)_ y\quad u'=\left[ \begin{matrix} a(\tau (y,t))\\ -b(\tau (y,t))+c(\tau (y,t))\end{matrix} \begin{matrix} b(\tau (y,t))-c(\tau (y,t))\\ - a(\tau (y,t))\end{matrix} \right]u\equiv A(\tau (y,t))u, \] \(y\in Y\), a compact metric space, and \(\tau\) : \(Y\times R\to Y\) is a flow in Y; a,b,c are continuous functions from Y into C. Associated with the equation \((1)_ y\) is the operator equation \[ L_ yu=\left[ \begin{matrix} 0\\ 1\end{matrix} \begin{matrix} -1\\ 0\end{matrix} \right][\frac{d}{dt}-A(\tau (y,t))]u=u,\quad y\in Y. \] The Floquet exponent w is defined using an ergodic measure \(\mu\) on Y. The concept of exponential dichotomy for equations \((1)_ y\) is basic in the authors' approach to the Floquet exponent. The remaining parts of the paper consider the definition of the holomorphic function \(w=w(\lambda)\) for complex \(\lambda\) and its close relation to the spectral properties of the operator \(L_ y\) and the Green's matrix (i.e. the kernel) \(G_ y(t,s;\lambda)\) of the operator \((\lambda -L_ y)^{-1}\), \(y\in Y\).