Maximal regularity for non-autonomous evolution equations governed by forms having less regularity

From MaRDI portal
Publication:2355908



Abstract: We consider the maximal regularity problem for non-autonomous evolution equations �egin{equation} left{ �egin{array}{rcl} u'(t) + A(t),u(t) &=& f(t), t in (0, au] u(0)&=&u_0. end{array} ight. end{equation} Each operator A(t) is associated with a sesquilinear form mathfraka(t) on a Hilbert space H. We assume that these forms all have the same domain V. It is proved in cite{HO14} that if the forms have some regularity with respect to t (e.g., piecewise alpha-H"older continuous for some alpha>1/2) then the above problem has maximal Lp--regularity for all u0 in the real-interpolation space (H,D(A(0)))1−1/p,p. In this paper we prove that the regularity required there can be improved for a class of sesquilinear forms. The forms considered here are such that the difference mathfraka(t;cdot,cdot)−mathfraka(s;cdot,cdot) is continuous on a larger space than the common domain V. We give three examples which illustrate our results.


The author considers the non-autonomous evolution equation \[ u'(t) + A(t)u(t) = f(t), u(0) = u_0. \] Here, each operator \(A(t)\) is associated to a form \(a(t,\cdot,\cdot)\) on some fixed Hilbert space \(H\). Moreover, the following standard assumptions are made: the form domains are independent of \(t\) and the forms are uniformly bounded and satisfy a uniform coercivity estimate in \(t\). A classical problem going back to Lions asks under which regularity assumptions on \(t \mapsto a(t,\cdot,\cdot)\) the problem has maximal \(L^p\)-regularity in \(H\). This means that for all \(f \in L^p(0,T;H)\) there exists a unique \(u \in W^{1,p}(0,T;H)\) with \(u(t) \in D(A(t))\) for almost all \(t\) that satisfies the equation in the \(L^p\)-sense. Several maximal regularity results for this problem have already been established by different authors. However, the minimal regularity on the forms needed is an open problem. In this note, the author proves a further sufficient criterion for maximal regularity using techniques developed in [\textit{B. Haak} and the author, ``Maximal regularity for non-autonomous evolution equations, Preprint, \url{arXiv:1402.1136}]: if the forms satisfy among the standard assumptions the regularity estimate \(| a(t,u,v) - a(s,u,v)| \leq \omega(| t - s|) \| u\|_{V_\beta}\| v\|_{V_\gamma}\) for some \(\beta\), \(\gamma \in [0,1]\) and for some non-decreasing \(\omega: [0,T] \to [0,\infty)\) with \[ \int_0^T \frac{\omega(t)}{t^{1+\frac{\gamma}{2}}} < \infty, \] then the problem has maximal \(L_p\)-regularity for \(u_0 = 0\). Here, \(V_\beta =[V,H]_\beta\) denotes the complex interpolation space between \(H\) and the common form domain \(V\). A similar criterion is shown for initial values in the real interpolation space \((H,D(A(0)))_{1-1/p,p}\). This generalizes recent results by \textit{W. Arendt} and {S. Monniaux} [``Maximal regularity for non-autonomous Robin boundary conditions, Preprint, \url{arXiv:1410.3063}]. At the end of the article, several concrete applications of the general theorem are given.




Cited in
(25)








This page was built for publication: Maximal regularity for non-autonomous evolution equations governed by forms having less regularity

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2355908)