Differential equations associated with compact evolution generators (Q2640777)

Let X be a real Banach space with norm \(\| \cdot \|\). Let J be the duality mapping of X and set \[ <y,x>_ i=\inf \{x^*(y),x^*\in J(x)\};\quad <y,x>_ -=<y,x>_ i\| x\|^{-1}\quad (1.1) \] \[ <y,x>_{\tilde s}=\sup \{x^*(y),x^*\in J(x)\};\quad <y,x>_+=<y,x>_{\tilde s}\| x\|^{-1} \] where y,x\(\in X\), \(x\neq 0.\) The paper studies the solution to the problem: \[ u'(t)\in A(t)u(t)+F(u(t)),\quad u(s)=u_ 0\in \overline{D(A(s))},\quad 0\leq t<T\quad (1.2) \] for some \(T>0.\) The assumptions on \(\{\) A(t), \(0\leq t\leq T\}\) and on F(t) are the following: (C1.2) \(\lim_{h\downarrow 0} h^{-1}d[x;R(I-hA(t+h)]=0\), \(x\in \overline{D(A(t))}\), \(0<t<T.\) (C2) There are two continuous functions f: [0,T], and L: [0,\(+\infty [\to [0,+\infty [\) such \(that:\) \(<y_ 1-y_ 2,x_ 1-x_ 2>_ i\leq \| f(t)-f(s)\| \| x_ 1- x_ 2\| L(\max \{\| x_ 1\|,\| x_ 2\| \})\) for all \(0\leq s\leq t\leq T\), \([x_ 1,y_ 1]\in A(t)\) and \([x_ 2,y_ 2]\in A(s).\) (C3) If \(t_ n\uparrow t\), \(x_ n\in D(A(t_ n))\) and \(x_ n\to x\), then \(x\in \overline{D(A(t))}\), \((t_ n,t\in [0,T]).\) (C4) The evolution operator U(t,s) generated by \(\{\) A(t), \(0\leq t\leq T\}\) is compact (i.e. for every \(0\leq s<t\leq T\), the operator \(x\to U(t,s)x\) maps bounded subsets of \(\overline{D(A(s))}\) into relatively compact (precompact) subset of X). The assumptions on F(t) are the following: (H1) The domain \(D(F(t))=\overline{D(A(t))}\), \(0\leq t<T.\) (H2) (t,x)\(\to F(t)x\) is continuous, i.e. if \(t_ n\to t\) and \(x_ n\to x\), \(x_ n\in D(F(t))\), then \(x\in D(F(t))\) and \(F(t_ n)\to F(t)x.\) The attention of the paper is devoted to the integral solution to \((1)+(2).\) Let g: [0,T]\(\to X\) a continuous function and \(s\in [0,T[\), \(s\leq T_ 1\leq T\). A function u: [s,T\({}_ 1]\to X\) is said to be an integral solution \(to\) u\({}'(t)\in A(t)u(t)+F(u(t))\), \(u(s)=u_ 0\in \overline{D(A(s))}\), \(0\leq t<T_ 1\) on \([s,T_ 1]\), if u is continuous on \([s,T_ 1]\), u(t)\(\in \overline{D(A(t))}\forall t\in [s,T_ 1]\), \(u(s)=u_ 0\) and the following inequality is \(satisfied\) \(\| u(t)-x\| \leq \| u(t_ 0)-x\| +\int^{t}_{0}(<y+g(t),u(\tau)-x>_++C\| f(\tau)-f(r)\|)d\tau\) for all \(s\leq t_ 0\leq t\leq T_ 1\), \(r\in [s,T_ 1]\), \(x\ni \overline{D(A(r))}\) and \(y\in A(r)x\), with \(C=L(\max \{\| u\|,\| x\| \})\), where \(\| u\| =\sup \{\| u(t)\|\), \(s\leq t\leq T_ 1\}.\) The main results of the paper are theorem 1 on the existence of an integral solution u(t) of (1.1), and a very interesting theorem 2 on the existence and behaviour of a maximal solution u(t).