Coupled intervals in the discrete calculus of variations: necessity and sufficiency (Q1856987)

As the authors state in the Introduction, ``in this paper a thorough study is presented of the nonnegativity and positivity of the discrete quadratic functional of the form'' \[ {\mathcal I}(\eta):=\eta^T_0\Gamma_0\eta_0 +\eta^T_{N+1}\Gamma\eta_{N+1} + \displaystyle \sum_{k=0}^N [\eta^T_{k+1}P_k\eta_{k+1}+2\eta^T_{k+1}Q_k\Delta \eta_k+\Delta\eta^T_kR_k\Delta\eta_k] \] over (variable) endpoints constraints of the form \[ M_0\eta_0=0, \;M\eta_{N+1}=0. \] In order to extend to the ``variable endpoints'' case some of their previous results [e.g., \textit{R. Hilscher} and \textit{V. Zeidan}, J. Differ. Equ. Appl. 8, 573--602 (2002; Zbl 1010.49020)], the authors introduce the new concept of coupled interval which is shown to contain as a particular case the classical concept of conjugate interval in calculus of variations. As in their previous papers, the authors prove a large number of results that contain different types of equivalent conditions for the nonnegativity and for the positivity of the functional \({\mathcal I}(\eta)\); these results are then used to obtain corresponding necessary and/or sufficient optimality conditions for local minima of some classes of problems in ``discrete'' calculus of variations.