Some second-order necessary conditions for nonconvex hyperbolic differential inclusion problems (Q5927587): Difference between revisions

It is considered the following problem: minimize \(g(u(T_1,T_2))\) with the constraints \(u_{xy}\in F(x,y,u(x,y))\), \(u_x(x,0)\in F_1(x,u(x,0))\), \(u_y(0,y)\in F_2(y,u(0,y))\), \(u(0,0)\in X_0\), \(u(T_1,T_2)\in X_1\), \(X_i\subset\mathbb R^n\); \(g\) is a locally-Lipschitz function; \(F\), \(F_i\) are compact-valued, nonconvex maps. The author proposes second-order optimality conditions. Clarke's nonsmooth analysis and some ideas from \textit{H. Zheng} [Appl. Math. Optimization 30, No. 1, 1-14 (1994; Zbl 0806.49017)] and \textit{A. Cernea} [in ``Diff. Inclus. and Optim. Contr.'', Lect. Notes Nonlinear Anal. 2, 113-121 (1998), per bibl.] have been used. It is claimed the possibility to replace \(F(x,y,u)\) by \(F(x,y,u,\nabla u)\).