Deferred correction methods for initial boundary value problems (Q1610532)

The numerical solution of the linear problem \[ {\partial u\over\partial t}={\mathcal P}(u)+ F(x,t),\quad u(x,0)= f(x) \] with appropriate boundary conditions by the method of lines gives rise to a large system of ordinary differential equations. The purpose of the paper is to apply the deferred correction principle to this system with initial conditions. The paper contains a proof of a stability estimate for a special case of time dependent coefficients as well as error estimates. It concludes with some numerical experiments.