Explicit solutions of higher order boundary value differential problems stable under data perturbations (Q1355034): Difference between revisions

The authors investigate solvability and dependence of solutions a the parameter for the linear boundary value problem \[ \begin{cases} x^{(p)}(t)+A_{p-1}(\lambda)x^{(p-1)}(t)+\dots+A_0(\lambda)x(t)=f(t),\quad t\in[0,b],\\ \sum_{j=0}^{p-1}\left\{E_{ij}(\lambda)x^{(j)}(0)+F_{ij}(\lambda)x^{(j)}(b)\right\}=G_i(\lambda),\quad 1\leq i\leq p,\end{cases}\tag{\(P(\lambda)\)} \] where \(x,f,G_i,i=1,\dots,p\) are vectors in \(\mathbb{C}^n\), \(A_j,E_{ij},F_{ij}\), \(0\leq j\leq p-1,\) \(1\leq i\leq p\), are \(n\times n\) matrices of continuously differentiable (with respect to \(\lambda\)) functions. In the first part of the paper an explicit formula for solution of (\(P(\lambda)\)) is presented. In this formula the solution \(x=x(t,\lambda)\) of (\(P(\lambda)\)) is expressed in terms of the Jordan normal form of the companion matrix \[ \begin{pmatrix} 0 &I &0 &\ldots &0\\ \vdots &\vdots &\vdots & &\vdots\\ 0 &0 &0 &\ldots &I\\ -A_0(\lambda) &-A_1(\lambda) &-A_2(\lambda)&\ldots &-A_{p-1}(\lambda)\end{pmatrix} \] and the matrices \(E_{ij}(\lambda),F_{ij(\lambda)}\) appearing in the boundary conditions. In the second part, given \(\lambda_0\) and \(\varepsilon>0\), the explicit value \(\delta>0\) (in terms of matrices appearing in (\(P(\lambda_0)\))) is found such that \(|\lambda-\lambda_0|\leq\delta\) implies that \(|v(t,\lambda)-x(t,\lambda_0)|\leq\varepsilon\) for \(t\in[0,b]\).