Second order nonlinear evolution inclusions. I: Existence and relaxation results (Q2581232): Difference between revisions

The authors study second-order nonlinear nonmonotone evolution inclusions defined in the framework of an evolution triple of spaces. In particular, they consider the problem \[ \ddot{x}(t) +A(t,\dot{x}(t))+Bx(t) \in F(t,x(t),\dot{x}(t)) \text{ a.e. }t \in T=[0,b],\;x(0)=z_0,\;\dot{x}(0)=z_1 , \] where \(A:T \times X \to X^*\) is a nonlinear operator, \(B:X \to X^*\) is a linear operator, \(F:T \times H \times H \to 2^H \backslash \{\emptyset \}\) is a multi-valued nonlinearity, \(z_0 \in X, z_1 \in H\) and \((X,H,X^*)\) is the evolution triple. In the case of an upper semicontinuous convex-valued nonlinearity, they prove that the set of the solutions is nonempty and compact in \(C^1(T,H)\) and they obtain also the existence of solutions when \(F\) is lower semicontinuous, not necessarily convex-valued. Moreover, they establish the existence of extremal solutions and they show that the set of such solutions is dense in \(C^1(T,H)\) to the set of the solutions of the convexified problem. In the last part, an example of a nonlinear hyperbolic optimal control problem is given.