Elliptic \(H^ 2\)-Volterra projection and the \(H^ 1\)-Galerkin methods for the integro-differential equations of evolution (Q1891673): Difference between revisions

Using ideas and techniques similar to those of \textit{J. R. Cannon} and \textit{Y. Lin} [SIAM J. Numer. Anal. 27, No. 3, 595-607 (1990; Zbl 0709.65122)], the author derives optimal \(L^2\), \(H^1\) and \(H^2\) error estimates for the \(H^1\)-Galerkin method applied to initial- boundary value problems for the parabolic integro-differential equation \[ u_t = \nabla \cdot (a(x) \nabla u + \int^t_0 b(x,t,\tau) \nabla u(x,\tau) d\tau) + f(u) \] and its hyperbolic counterpart. This is done via the elliptic \(H^2\)-Volterra projection; the error estimates are obtained for the semidiscrete and the fully discrete \(H^1\)- Galerkin methods.