Piecewise spectral collocation method for system of Volterra integral equations (Q2400486)

This paper describes a collocation method for the numerical solution of a system of linear Volterra integral equations of the form \[ {\mathbf y}(t) = {\mathbf g}(t) + \int_0^t {\mathbf K}(t,s) {\mathbf y}(s) \,ds. \] The interval \([0,T]\) is divided into subintervals of width \(\leq h\). On each subinterval, \(N+1\) Gauss-Lobatto points are used as collocation points. Most of the paper is taken up with the proof of the main theorem, which states that if \({\mathbf g}\), \({\mathbf K}\) have \(m\) continuous derivatives, the error satisfies \[ \| {\mathbf e} \|_\infty \leq C \cdot h^{m-1/2} N^{1/2-m}, \] \[ \| {\mathbf e} \|_2 \leq C \cdot h^{m} N^{-m}. \] The result is verified with some numerical examples.