Error bounds for the solution of Volterra and delay equations (Q1189137)

The construction of efficient and reliable numerical software requires a theory for two distinct issues: a convergence analysis and a way of bounding the error in the computed results. While the first of these questions has received much attention, less is known about the second one. The authors describe a method for bounding the error in the solution of Fredholm integral equations and show how the approach can be extended to Volterra and delay equations.