Estimation of time dependent parameters in general parabolic evolution systems (Q1815456)

The authors present a unified theoretical and computational approach for estimating time dependent parameters in abstract parabolic systems. Results are proved initially, within the framework of a Gelfand triple \(V\hookrightarrow H\hookrightarrow V^*\), for the abstract non-autonomous equation \[ \dot u(t,q)= A(t,q)u(t,q)+ F(t,u(t,q),q),\quad u(0,q)= u_0(q).\tag{1} \] In (1) it is assumed that the parameter \(q\) belongs to a compact separable metric space, the linear operator \(A(t,q)\) is determined by a (suitably restricted) time and parameter dependent sesquilinear form on \(V\), and the nonlinear term \(F\) satisfies certain continuity conditions. The existence and uniqueness of a solution to a weak formulation of (1) is established and a convergence theory for least squares based parameter estimation is produced. The general theory is then applied to the specific cases of a contaminated groundwater model and the Euler-Bernoulli beam equation.