Local-Cubic Spline for Approximate Solution of Boundary Value Problems

We have constructed an explicit local-cubic spline for the approximation of the smooth functions and have studied the behavior of the approximation. To solve numerically boundary value problems, a spline-scheme based on the properties of the local-cubic spline and the standard cubic spline collocation is proposed. The scheme is implemented by sequentially solving two tridiagonal systems, which allow to use the three-point sweep method and differ from each other only by matrix of the right-hand side of the equation. It indicates that this algorithm is efficient. The number of operations depends linearly on the number of grid nodes. It is proved that the constructed spline possesses the same approximation properties as the local-cubic spline. Thus, in this paper we actually considered the approximation of the solutions of the boundary value problems. The proposed scheme also allows to find the first and second derivatives of the solution of the boundary value problem on the uniform grid nodes of the fourth-order accuracy with respect to the step-size of the grid. The numerical experiments confirm the theoretical order of convergence. Due to good approximation properties and the simplicity of the algorithm implementation, the proposed method can be applied to solve numerically the boundary value problems for the second order ordinary differential equations, which often occur in mathematics, physics, and in the field of natural and engineering sciences.

boundary value problems, cubic spline, high accuracy