Topological theory of non-autonomous parabolic evolution inclusions on a noncompact interval and applications (Q2354324)

The authors study the semilinear evolution inclusion \[ u^{\prime}(t)-A(t)u(t)\in F(t,u(t),u_{t}),\;t\in [ 0,\infty )\tag{1} \] in a real Banach space \(X\), subject to either the condition \[ u(t)=\phi (t),\;t\in [ -\tau ,0]\tag{2} \] or \[ u(t)=H(u)(t),\;t\in [ -\tau ,0],\tag{3} \] where \(\left( A(t),D(A(t))\right) \) is a family of (possibly unbounded) linear operators on \(X\), \(F:[0,\infty )\times X\times C\left( \left[ -\tau ,0\right] ;X\right) \rightarrow 2^{X}\) is convex- and closed-valued, \(u_{t}\in C\left( \left[ -\tau ,0\right] ;X\right) \) where \(u_{t}(s)\equiv u(t+s)\) for \(s\in \left[ -\tau ,0\right] \) and \(u\in \tilde{C}\left( [-\tau ,\infty );X\right) \), \(\phi \in C\left( \left[ -\tau ,0\right] ;X\right) \) and \(H:\tilde{C}_{b}\left( [-\tau ,\infty );X\right) \rightarrow C\left( \left[ -\tau ,0\right] ;X\right) \), where \(\tilde{C}\left( [-\tau ,\infty );X\right) \) and \(\tilde{C}_{b}\left( [-\tau ,\infty );X\right) \) are certain spaces of continuous functions. It is assumed that \(F(t,\cdot ,\cdot )\) is weakly upper semicontinuous and that \(F(\cdot ,x,v)\) has a strongly measurable selection for each \((x,v)\in X\times C\left( \left[ -\tau ,0\right] ;X\right) \), and the family \(A(t)\) satisfies the ``Acquistapace-Terrini conditions''. It is proven that the solution set is a compact \(R_{\delta }\) set for problem (1), (2) using the ``inverse limit method'' from [\textit{J. Andres} and \textit{M. Pavlačková}, J. Differ. Equations 248, No. 1, 127--150 (2010; Zbl 1188.34015)] and [\textit{D.-H. Chen} et al., J. Funct. Anal. 265, No. 9, 2039--2073 (2013; Zbl 1287.34055)]. \ (A subset \(S\) of a metric space is an \(R_{\delta }\) set if there exists a nondecreasing sequence \(\left\{ S_{n}\right\} \) of compact and contractible sets such that \(S=\bigcap_{n=1}^{\infty }S_{n}\).) The existence of global integral solutions to the problem (1), (3) is proven using the previous result. The authors note that no nonexpansivity condition is made on \(H\), unlike some previous results. The first theorem is extended to a control problem, which in turn is used to study the invariance of the reachability set of the control problem under single-valued nonlinear perturbations. Finally, two nontrivial examples are given.