Two generalized classes of exponentially dichotomous linear differential systems on the time half-line without uniform estimates for the solution norms. I (Q2188049)

The article is focused on the properties of weak exponential dichotomy and almost exponential dichotomy for the family \(\mathcal{M}_{n}\) of linear ODE systems \[ \dot{x}=A(t)x \quad x\in \mathbb{R}^{n}, \quad t\geq 0, \] where the coefficients \(t\mapsto a_{ij}(t)\) are piecewise continuous and bounded on the half-line \(t\geq 0\). More specifically, the article has a double contribution: a) provides a characterization of the weak exponential dichotomy in terms of Bohl exponents and b) obtains some relations between these dichotomies and the classical exponetial dichotomy for special families. A system in \(\mathcal{M}_{n}\) is said to be weakly exponentially dichotomous on the half line if there exists positive constants \(\nu_{1}\) and \(\nu_{2}\) and a decomposition of the space \(\mathbb{R}^{n}\) of initial data (at \(t=0\)) into a direct sum of subspaces \(L_{-}\) and \(L_{+}\) (the case of zero dimension of one of the subspaces not being excluded) such that the following conditions are satisfied: \begin{itemize} \item[1'.] If \(x(0)\in L_{-}\), then \(||x(t)||\leq c_{1}(x)e^{-\nu_{1}(t-s)}||x(s)|| \,\forall t\geq s \geq 0\). \item[2'.] If \(x(0)\in L_{+}\), then \(||x(t)||\geq c_{2}(x)e^{\nu_{2}(t-s)}||x(s)|| \, \forall t\geq s \geq 0\), \end{itemize} where \(c_{1}(x)\) and \(c_{2}(x)\) are positive constants generally depending on the choice of the solution \(x(\cdot)\). Moreover, a system in \(\mathcal{M}_{n}\) is said to be almost exponentially dichotomous on the half line if there exists positive constants \(c_{1}\),\(c_{2}\),\(\nu_{1}\) and \(\nu_{2}\) and a decomposition of the space \(\mathbb{R}^{n}\) of initial data (at \(t=0\)) into a direct sum of subspaces \(L_{-}\) and \(L_{+}\) (the case of zero dimension of one of the subspaces not being excluded) such that the following conditions are satisfied \begin{itemize} \item[1''.] If \(x(0)\in L_{-}\), then \(||x(t)||\leq c_{1}e^{-\nu_{1}(t-s)}||x(s)|| \, \forall t\geq s \geq t_{x}\). \item[2''.] If \(x(0)\in L_{+}\), then \(||x(t)||\geq c_{2}e^{\nu_{2}(t-s)}||x(s)|| \,\forall t\geq s \geq t_{x}\), \end{itemize} where \(t_{x}\) is a nonnegative number generally depending on the choice of the solution \(x(\cdot)\). The families of systems weakly exponentially autonomous and almost exponentially autonomous are respectively denoted by \(W\mathcal{E}_{n}\) and \(A\mathcal{E}_{n}\). Notice that, if \(c_{1}(x)=c_{1}\) and \(c_{2}=c_{2}\) or \(t_{x}=0\) for any solution \(x(\cdot)\), then the classical exponential dichotomy on the half line is recovered and will be denoted by \(\mathcal{E}_{n}\). In addition, \(A\mathcal{E}_{n}^{m}\) and \(W\mathcal{E}_{n}^{m}\) are the subclasses of \(A\mathcal{E}_{n}\) and \(W\mathcal{E}_{n}\) consisting in systems that have the dimension of their subspace \(L_{-}\) equal to \(m\) with \(0\leq m \leq n\). The first main result (Theorem 1) provides a necessary and sufficient condition ensuring that a system in \(\mathcal{M}_{n}\) is weakly exponentially dichotomous on the half line. This condition is described in terms of upper and lower Bohl exponents. The second main result (Theorem 2) states that for the cases \((n,m)=(1,0)\), \((n,m)=(1,1)\) and \((n,m)=(2,1)\), it follows that \(\mathcal{E}_{n}^{m}=A\mathcal{E}_{n}^{m}=W\mathcal{E}_{n}^{m}\). For the remainig pairs with \(n\in \mathbb{N}\) and \(0\leq m\leq n\), it is proved that \[ \mathcal{E}_{n}^{m} \subset A\mathcal{E}_{n}^{m} \subset W\mathcal{E}_{n}^{m}. \]