High order linear initial-value problems and Schauder bases (Q2471262)

A new numerical method is proposed for solving the initial value problem defined by the non-homogeneous, variable coefficient matrix, \(m\)th-order, linear, ordinary differential equation. It is based on the geometric series theorem and the use of the classical Schauder basis for a certain Banach space associated with the problem. With these tools, a sequence of functions \(\{ y_r \}_{r \geq 0}\) is constructed in such a way that the difference in norm with the usual approximations provided by a Picard-like iteration process is small enough. An important feature of the procedure is that the approximations \(\{ y_r \}_{r \geq 0}\) are easily obtained as linear iterative expressions, in contrast with classical methods. The main properties of the method are analyzed and a pair of numerical examples involving \(3 \times 3\) matrices are considered. The second one illustrates the numerical simulation of the response of a structural model to an external excitation, of interest in civil engineering, but no comparison with other numerical schemes is reported in the paper.